[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/******************************************************************************
|
2009-08-12 12:04:37 +00:00
|
|
|
* emulate.c
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
*
|
|
|
|
|
* Generic x86 (32-bit and 64-bit) instruction decoder and emulator.
|
|
|
|
|
*
|
|
|
|
|
* Copyright (c) 2005 Keir Fraser
|
|
|
|
|
*
|
|
|
|
|
* Linux coding style, mod r/m decoder, segment base fixes, real-mode
|
2007-07-17 13:16:56 +00:00
|
|
|
* privileged instructions:
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
*
|
|
|
|
|
* Copyright (C) 2006 Qumranet
|
2010-10-06 12:23:22 +00:00
|
|
|
* Copyright 2010 Red Hat, Inc. and/or its affiliates.
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
*
|
|
|
|
|
* Avi Kivity <avi@qumranet.com>
|
|
|
|
|
* Yaniv Kamay <yaniv@qumranet.com>
|
|
|
|
|
*
|
|
|
|
|
* This work is licensed under the terms of the GNU GPL, version 2. See
|
|
|
|
|
* the COPYING file in the top-level directory.
|
|
|
|
|
*
|
|
|
|
|
* From: xen-unstable 10676:af9809f51f81a3c43f276f00c81a52ef558afda4
|
|
|
|
|
*/
|
|
|
|
|
|
2007-12-16 09:02:48 +00:00
|
|
|
#include <linux/kvm_host.h>
|
2008-06-27 17:58:02 +00:00
|
|
|
#include "kvm_cache_regs.h"
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
#include <linux/module.h>
|
2009-08-12 12:04:37 +00:00
|
|
|
#include <asm/kvm_emulate.h>
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2010-01-21 13:31:48 +00:00
|
|
|
#include "x86.h"
|
2010-03-18 13:20:17 +00:00
|
|
|
#include "tss.h"
|
2009-06-17 13:50:33 +00:00
|
|
|
|
2011-09-13 07:45:41 +00:00
|
|
|
/*
|
|
|
|
|
* Operand types
|
|
|
|
|
*/
|
2011-09-13 07:45:42 +00:00
|
|
|
#define OpNone 0ull
|
|
|
|
|
#define OpImplicit 1ull /* No generic decode */
|
|
|
|
|
#define OpReg 2ull /* Register */
|
|
|
|
|
#define OpMem 3ull /* Memory */
|
|
|
|
|
#define OpAcc 4ull /* Accumulator: AL/AX/EAX/RAX */
|
|
|
|
|
#define OpDI 5ull /* ES:DI/EDI/RDI */
|
|
|
|
|
#define OpMem64 6ull /* Memory, 64-bit */
|
|
|
|
|
#define OpImmUByte 7ull /* Zero-extended 8-bit immediate */
|
|
|
|
|
#define OpDX 8ull /* DX register */
|
2011-09-13 07:45:43 +00:00
|
|
|
#define OpCL 9ull /* CL register (for shifts) */
|
|
|
|
|
#define OpImmByte 10ull /* 8-bit sign extended immediate */
|
|
|
|
|
#define OpOne 11ull /* Implied 1 */
|
|
|
|
|
#define OpImm 12ull /* Sign extended immediate */
|
2011-09-13 07:45:47 +00:00
|
|
|
#define OpMem16 13ull /* Memory operand (16-bit). */
|
|
|
|
|
#define OpMem32 14ull /* Memory operand (32-bit). */
|
|
|
|
|
#define OpImmU 15ull /* Immediate operand, zero extended */
|
|
|
|
|
#define OpSI 16ull /* SI/ESI/RSI */
|
|
|
|
|
#define OpImmFAddr 17ull /* Immediate far address */
|
|
|
|
|
#define OpMemFAddr 18ull /* Far address in memory */
|
|
|
|
|
#define OpImmU16 19ull /* Immediate operand, 16 bits, zero extended */
|
2011-09-13 07:45:49 +00:00
|
|
|
#define OpES 20ull /* ES */
|
|
|
|
|
#define OpCS 21ull /* CS */
|
|
|
|
|
#define OpSS 22ull /* SS */
|
|
|
|
|
#define OpDS 23ull /* DS */
|
|
|
|
|
#define OpFS 24ull /* FS */
|
|
|
|
|
#define OpGS 25ull /* GS */
|
2012-01-16 13:08:44 +00:00
|
|
|
#define OpMem8 26ull /* 8-bit zero extended memory operand */
|
2011-09-13 07:45:47 +00:00
|
|
|
|
|
|
|
|
#define OpBits 5 /* Width of operand field */
|
2011-09-13 07:45:42 +00:00
|
|
|
#define OpMask ((1ull << OpBits) - 1)
|
2011-09-13 07:45:41 +00:00
|
|
|
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/*
|
|
|
|
|
* Opcode effective-address decode tables.
|
|
|
|
|
* Note that we only emulate instructions that have at least one memory
|
|
|
|
|
* operand (excluding implicit stack references). We assume that stack
|
|
|
|
|
* references and instruction fetches will never occur in special memory
|
|
|
|
|
* areas that require emulation. So, for example, 'mov <imm>,<reg>' need
|
|
|
|
|
* not be handled.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/* Operand sizes: 8-bit operands or specified/overridden size. */
|
2010-07-29 12:11:49 +00:00
|
|
|
#define ByteOp (1<<0) /* 8-bit operands. */
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/* Destination operand type. */
|
2011-09-13 07:45:41 +00:00
|
|
|
#define DstShift 1
|
|
|
|
|
#define ImplicitOps (OpImplicit << DstShift)
|
|
|
|
|
#define DstReg (OpReg << DstShift)
|
|
|
|
|
#define DstMem (OpMem << DstShift)
|
|
|
|
|
#define DstAcc (OpAcc << DstShift)
|
|
|
|
|
#define DstDI (OpDI << DstShift)
|
|
|
|
|
#define DstMem64 (OpMem64 << DstShift)
|
|
|
|
|
#define DstImmUByte (OpImmUByte << DstShift)
|
|
|
|
|
#define DstDX (OpDX << DstShift)
|
|
|
|
|
#define DstMask (OpMask << DstShift)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/* Source operand type. */
|
2011-09-13 07:45:47 +00:00
|
|
|
#define SrcShift 6
|
|
|
|
|
#define SrcNone (OpNone << SrcShift)
|
|
|
|
|
#define SrcReg (OpReg << SrcShift)
|
|
|
|
|
#define SrcMem (OpMem << SrcShift)
|
|
|
|
|
#define SrcMem16 (OpMem16 << SrcShift)
|
|
|
|
|
#define SrcMem32 (OpMem32 << SrcShift)
|
|
|
|
|
#define SrcImm (OpImm << SrcShift)
|
|
|
|
|
#define SrcImmByte (OpImmByte << SrcShift)
|
|
|
|
|
#define SrcOne (OpOne << SrcShift)
|
|
|
|
|
#define SrcImmUByte (OpImmUByte << SrcShift)
|
|
|
|
|
#define SrcImmU (OpImmU << SrcShift)
|
|
|
|
|
#define SrcSI (OpSI << SrcShift)
|
|
|
|
|
#define SrcImmFAddr (OpImmFAddr << SrcShift)
|
|
|
|
|
#define SrcMemFAddr (OpMemFAddr << SrcShift)
|
|
|
|
|
#define SrcAcc (OpAcc << SrcShift)
|
|
|
|
|
#define SrcImmU16 (OpImmU16 << SrcShift)
|
|
|
|
|
#define SrcDX (OpDX << SrcShift)
|
2012-01-16 13:08:44 +00:00
|
|
|
#define SrcMem8 (OpMem8 << SrcShift)
|
2011-09-13 07:45:47 +00:00
|
|
|
#define SrcMask (OpMask << SrcShift)
|
2011-05-30 18:23:14 +00:00
|
|
|
#define BitOp (1<<11)
|
|
|
|
|
#define MemAbs (1<<12) /* Memory operand is absolute displacement */
|
|
|
|
|
#define String (1<<13) /* String instruction (rep capable) */
|
|
|
|
|
#define Stack (1<<14) /* Stack instruction (push/pop) */
|
|
|
|
|
#define GroupMask (7<<15) /* Opcode uses one of the group mechanisms */
|
|
|
|
|
#define Group (1<<15) /* Bits 3:5 of modrm byte extend opcode */
|
|
|
|
|
#define GroupDual (2<<15) /* Alternate decoding of mod == 3 */
|
|
|
|
|
#define Prefix (3<<15) /* Instruction varies with 66/f2/f3 prefix */
|
|
|
|
|
#define RMExt (4<<15) /* Opcode extension in ModRM r/m if mod == 3 */
|
|
|
|
|
#define Sse (1<<18) /* SSE Vector instruction */
|
2011-09-13 07:45:44 +00:00
|
|
|
/* Generic ModRM decode. */
|
|
|
|
|
#define ModRM (1<<19)
|
|
|
|
|
/* Destination is only written; never read. */
|
|
|
|
|
#define Mov (1<<20)
|
2009-08-23 11:24:25 +00:00
|
|
|
/* Misc flags */
|
2011-04-04 10:39:26 +00:00
|
|
|
#define Prot (1<<21) /* instruction generates #UD if not in prot-mode */
|
2011-02-01 14:32:03 +00:00
|
|
|
#define VendorSpecific (1<<22) /* Vendor specific instruction */
|
2010-08-01 12:10:29 +00:00
|
|
|
#define NoAccess (1<<23) /* Don't access memory (lea/invlpg/verr etc) */
|
2010-08-01 11:46:54 +00:00
|
|
|
#define Op3264 (1<<24) /* Operand is 64b in long mode, 32b otherwise */
|
2010-07-26 11:37:47 +00:00
|
|
|
#define Undefined (1<<25) /* No Such Instruction */
|
2010-02-10 12:21:36 +00:00
|
|
|
#define Lock (1<<26) /* lock prefix is allowed for the instruction */
|
2010-02-10 12:21:35 +00:00
|
|
|
#define Priv (1<<27) /* instruction generates #GP if current CPL != 0 */
|
2009-08-23 11:24:25 +00:00
|
|
|
#define No64 (1<<28)
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
#define PageTable (1 << 29) /* instruction used to write page table */
|
2008-12-04 13:26:42 +00:00
|
|
|
/* Source 2 operand type */
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
#define Src2Shift (30)
|
2011-09-13 07:45:43 +00:00
|
|
|
#define Src2None (OpNone << Src2Shift)
|
|
|
|
|
#define Src2CL (OpCL << Src2Shift)
|
|
|
|
|
#define Src2ImmByte (OpImmByte << Src2Shift)
|
|
|
|
|
#define Src2One (OpOne << Src2Shift)
|
|
|
|
|
#define Src2Imm (OpImm << Src2Shift)
|
2011-09-13 07:45:49 +00:00
|
|
|
#define Src2ES (OpES << Src2Shift)
|
|
|
|
|
#define Src2CS (OpCS << Src2Shift)
|
|
|
|
|
#define Src2SS (OpSS << Src2Shift)
|
|
|
|
|
#define Src2DS (OpDS << Src2Shift)
|
|
|
|
|
#define Src2FS (OpFS << Src2Shift)
|
|
|
|
|
#define Src2GS (OpGS << Src2Shift)
|
2011-09-13 07:45:43 +00:00
|
|
|
#define Src2Mask (OpMask << Src2Shift)
|
2012-04-09 15:40:02 +00:00
|
|
|
#define Mmx ((u64)1 << 40) /* MMX Vector instruction */
|
2012-04-09 15:39:59 +00:00
|
|
|
#define Aligned ((u64)1 << 41) /* Explicitly aligned (e.g. MOVDQA) */
|
|
|
|
|
#define Unaligned ((u64)1 << 42) /* Explicitly unaligned (e.g. MOVDQU) */
|
|
|
|
|
#define Avx ((u64)1 << 43) /* Advanced Vector Extensions */
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2010-07-29 12:11:54 +00:00
|
|
|
#define X2(x...) x, x
|
|
|
|
|
#define X3(x...) X2(x), x
|
|
|
|
|
#define X4(x...) X2(x), X2(x)
|
|
|
|
|
#define X5(x...) X4(x), x
|
|
|
|
|
#define X6(x...) X4(x), X2(x)
|
|
|
|
|
#define X7(x...) X4(x), X3(x)
|
|
|
|
|
#define X8(x...) X4(x), X4(x)
|
|
|
|
|
#define X16(x...) X8(x), X8(x)
|
2010-07-26 11:37:39 +00:00
|
|
|
|
2010-07-29 12:11:35 +00:00
|
|
|
struct opcode {
|
2011-09-13 07:45:42 +00:00
|
|
|
u64 flags : 56;
|
|
|
|
|
u64 intercept : 8;
|
2010-07-29 12:11:39 +00:00
|
|
|
union {
|
2010-07-29 12:11:51 +00:00
|
|
|
int (*execute)(struct x86_emulate_ctxt *ctxt);
|
2010-07-29 12:11:39 +00:00
|
|
|
struct opcode *group;
|
|
|
|
|
struct group_dual *gdual;
|
2011-03-29 09:34:38 +00:00
|
|
|
struct gprefix *gprefix;
|
2010-07-29 12:11:39 +00:00
|
|
|
} u;
|
2011-04-04 10:39:25 +00:00
|
|
|
int (*check_perm)(struct x86_emulate_ctxt *ctxt);
|
2010-07-29 12:11:39 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
struct group_dual {
|
|
|
|
|
struct opcode mod012[8];
|
|
|
|
|
struct opcode mod3[8];
|
2010-07-29 12:11:35 +00:00
|
|
|
};
|
|
|
|
|
|
2011-03-29 09:34:38 +00:00
|
|
|
struct gprefix {
|
|
|
|
|
struct opcode pfx_no;
|
|
|
|
|
struct opcode pfx_66;
|
|
|
|
|
struct opcode pfx_f2;
|
|
|
|
|
struct opcode pfx_f3;
|
|
|
|
|
};
|
|
|
|
|
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/* EFLAGS bit definitions. */
|
2010-02-10 12:21:34 +00:00
|
|
|
#define EFLG_ID (1<<21)
|
|
|
|
|
#define EFLG_VIP (1<<20)
|
|
|
|
|
#define EFLG_VIF (1<<19)
|
|
|
|
|
#define EFLG_AC (1<<18)
|
2009-06-17 13:50:32 +00:00
|
|
|
#define EFLG_VM (1<<17)
|
|
|
|
|
#define EFLG_RF (1<<16)
|
2010-02-10 12:21:34 +00:00
|
|
|
#define EFLG_IOPL (3<<12)
|
|
|
|
|
#define EFLG_NT (1<<14)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
#define EFLG_OF (1<<11)
|
|
|
|
|
#define EFLG_DF (1<<10)
|
2009-06-17 13:50:32 +00:00
|
|
|
#define EFLG_IF (1<<9)
|
2010-02-10 12:21:34 +00:00
|
|
|
#define EFLG_TF (1<<8)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
#define EFLG_SF (1<<7)
|
|
|
|
|
#define EFLG_ZF (1<<6)
|
|
|
|
|
#define EFLG_AF (1<<4)
|
|
|
|
|
#define EFLG_PF (1<<2)
|
|
|
|
|
#define EFLG_CF (1<<0)
|
|
|
|
|
|
2010-07-28 09:38:40 +00:00
|
|
|
#define EFLG_RESERVED_ZEROS_MASK 0xffc0802a
|
|
|
|
|
#define EFLG_RESERVED_ONE_MASK 2
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
static ulong reg_read(struct x86_emulate_ctxt *ctxt, unsigned nr)
|
|
|
|
|
{
|
|
|
|
|
if (!(ctxt->regs_valid & (1 << nr))) {
|
|
|
|
|
ctxt->regs_valid |= 1 << nr;
|
|
|
|
|
ctxt->_regs[nr] = ctxt->ops->read_gpr(ctxt, nr);
|
|
|
|
|
}
|
|
|
|
|
return ctxt->_regs[nr];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static ulong *reg_write(struct x86_emulate_ctxt *ctxt, unsigned nr)
|
|
|
|
|
{
|
|
|
|
|
ctxt->regs_valid |= 1 << nr;
|
|
|
|
|
ctxt->regs_dirty |= 1 << nr;
|
|
|
|
|
return &ctxt->_regs[nr];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static ulong *reg_rmw(struct x86_emulate_ctxt *ctxt, unsigned nr)
|
|
|
|
|
{
|
|
|
|
|
reg_read(ctxt, nr);
|
|
|
|
|
return reg_write(ctxt, nr);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void writeback_registers(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
unsigned reg;
|
|
|
|
|
|
|
|
|
|
for_each_set_bit(reg, (ulong *)&ctxt->regs_dirty, 16)
|
|
|
|
|
ctxt->ops->write_gpr(ctxt, reg, ctxt->_regs[reg]);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void invalidate_registers(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
ctxt->regs_dirty = 0;
|
|
|
|
|
ctxt->regs_valid = 0;
|
|
|
|
|
}
|
|
|
|
|
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/*
|
|
|
|
|
* Instruction emulation:
|
|
|
|
|
* Most instructions are emulated directly via a fragment of inline assembly
|
|
|
|
|
* code. This allows us to save/restore EFLAGS and thus very easily pick up
|
|
|
|
|
* any modified flags.
|
|
|
|
|
*/
|
|
|
|
|
|
2006-12-13 08:33:45 +00:00
|
|
|
#if defined(CONFIG_X86_64)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
#define _LO32 "k" /* force 32-bit operand */
|
|
|
|
|
#define _STK "%%rsp" /* stack pointer */
|
|
|
|
|
#elif defined(__i386__)
|
|
|
|
|
#define _LO32 "" /* force 32-bit operand */
|
|
|
|
|
#define _STK "%%esp" /* stack pointer */
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* These EFLAGS bits are restored from saved value during emulation, and
|
|
|
|
|
* any changes are written back to the saved value after emulation.
|
|
|
|
|
*/
|
|
|
|
|
#define EFLAGS_MASK (EFLG_OF|EFLG_SF|EFLG_ZF|EFLG_AF|EFLG_PF|EFLG_CF)
|
|
|
|
|
|
|
|
|
|
/* Before executing instruction: restore necessary bits in EFLAGS. */
|
2007-12-06 14:15:02 +00:00
|
|
|
#define _PRE_EFLAGS(_sav, _msk, _tmp) \
|
|
|
|
|
/* EFLAGS = (_sav & _msk) | (EFLAGS & ~_msk); _sav &= ~_msk; */ \
|
|
|
|
|
"movl %"_sav",%"_LO32 _tmp"; " \
|
|
|
|
|
"push %"_tmp"; " \
|
|
|
|
|
"push %"_tmp"; " \
|
|
|
|
|
"movl %"_msk",%"_LO32 _tmp"; " \
|
|
|
|
|
"andl %"_LO32 _tmp",("_STK"); " \
|
|
|
|
|
"pushf; " \
|
|
|
|
|
"notl %"_LO32 _tmp"; " \
|
|
|
|
|
"andl %"_LO32 _tmp",("_STK"); " \
|
|
|
|
|
"andl %"_LO32 _tmp","__stringify(BITS_PER_LONG/4)"("_STK"); " \
|
|
|
|
|
"pop %"_tmp"; " \
|
|
|
|
|
"orl %"_LO32 _tmp",("_STK"); " \
|
|
|
|
|
"popf; " \
|
|
|
|
|
"pop %"_sav"; "
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
|
|
|
|
/* After executing instruction: write-back necessary bits in EFLAGS. */
|
|
|
|
|
#define _POST_EFLAGS(_sav, _msk, _tmp) \
|
|
|
|
|
/* _sav |= EFLAGS & _msk; */ \
|
|
|
|
|
"pushf; " \
|
|
|
|
|
"pop %"_tmp"; " \
|
|
|
|
|
"andl %"_msk",%"_LO32 _tmp"; " \
|
|
|
|
|
"orl %"_LO32 _tmp",%"_sav"; "
|
|
|
|
|
|
2008-11-26 13:14:10 +00:00
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
#define ON64(x) x
|
|
|
|
|
#else
|
|
|
|
|
#define ON64(x)
|
|
|
|
|
#endif
|
|
|
|
|
|
2011-09-07 13:41:35 +00:00
|
|
|
#define ____emulate_2op(ctxt, _op, _x, _y, _suffix, _dsttype) \
|
2008-11-26 13:30:45 +00:00
|
|
|
do { \
|
|
|
|
|
__asm__ __volatile__ ( \
|
|
|
|
|
_PRE_EFLAGS("0", "4", "2") \
|
|
|
|
|
_op _suffix " %"_x"3,%1; " \
|
|
|
|
|
_POST_EFLAGS("0", "4", "2") \
|
2011-09-07 13:41:35 +00:00
|
|
|
: "=m" ((ctxt)->eflags), \
|
|
|
|
|
"+q" (*(_dsttype*)&(ctxt)->dst.val), \
|
2008-11-26 13:30:45 +00:00
|
|
|
"=&r" (_tmp) \
|
2011-09-07 13:41:35 +00:00
|
|
|
: _y ((ctxt)->src.val), "i" (EFLAGS_MASK)); \
|
2008-11-29 18:38:12 +00:00
|
|
|
} while (0)
|
2008-11-26 13:30:45 +00:00
|
|
|
|
|
|
|
|
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/* Raw emulation: instruction has two explicit operands. */
|
2011-09-07 13:41:35 +00:00
|
|
|
#define __emulate_2op_nobyte(ctxt,_op,_wx,_wy,_lx,_ly,_qx,_qy) \
|
2008-11-26 13:30:45 +00:00
|
|
|
do { \
|
|
|
|
|
unsigned long _tmp; \
|
|
|
|
|
\
|
2011-09-07 13:41:35 +00:00
|
|
|
switch ((ctxt)->dst.bytes) { \
|
2008-11-26 13:30:45 +00:00
|
|
|
case 2: \
|
2011-09-07 13:41:35 +00:00
|
|
|
____emulate_2op(ctxt,_op,_wx,_wy,"w",u16); \
|
2008-11-26 13:30:45 +00:00
|
|
|
break; \
|
|
|
|
|
case 4: \
|
2011-09-07 13:41:35 +00:00
|
|
|
____emulate_2op(ctxt,_op,_lx,_ly,"l",u32); \
|
2008-11-26 13:30:45 +00:00
|
|
|
break; \
|
|
|
|
|
case 8: \
|
2011-09-07 13:41:35 +00:00
|
|
|
ON64(____emulate_2op(ctxt,_op,_qx,_qy,"q",u64)); \
|
2008-11-26 13:30:45 +00:00
|
|
|
break; \
|
|
|
|
|
} \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
2011-09-07 13:41:35 +00:00
|
|
|
#define __emulate_2op(ctxt,_op,_bx,_by,_wx,_wy,_lx,_ly,_qx,_qy) \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
do { \
|
2008-11-26 13:30:45 +00:00
|
|
|
unsigned long _tmp; \
|
2011-09-07 13:41:35 +00:00
|
|
|
switch ((ctxt)->dst.bytes) { \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
case 1: \
|
2011-09-07 13:41:35 +00:00
|
|
|
____emulate_2op(ctxt,_op,_bx,_by,"b",u8); \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break; \
|
|
|
|
|
default: \
|
2011-09-07 13:41:35 +00:00
|
|
|
__emulate_2op_nobyte(ctxt, _op, \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
_wx, _wy, _lx, _ly, _qx, _qy); \
|
|
|
|
|
break; \
|
|
|
|
|
} \
|
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
|
|
/* Source operand is byte-sized and may be restricted to just %cl. */
|
2011-09-07 13:41:35 +00:00
|
|
|
#define emulate_2op_SrcB(ctxt, _op) \
|
|
|
|
|
__emulate_2op(ctxt, _op, "b", "c", "b", "c", "b", "c", "b", "c")
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
|
|
|
|
/* Source operand is byte, word, long or quad sized. */
|
2011-09-07 13:41:35 +00:00
|
|
|
#define emulate_2op_SrcV(ctxt, _op) \
|
|
|
|
|
__emulate_2op(ctxt, _op, "b", "q", "w", "r", _LO32, "r", "", "r")
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
|
|
|
|
/* Source operand is word, long or quad sized. */
|
2011-09-07 13:41:35 +00:00
|
|
|
#define emulate_2op_SrcV_nobyte(ctxt, _op) \
|
|
|
|
|
__emulate_2op_nobyte(ctxt, _op, "w", "r", _LO32, "r", "", "r")
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2008-12-04 13:29:00 +00:00
|
|
|
/* Instruction has three operands and one operand is stored in ECX register */
|
2011-09-07 13:41:37 +00:00
|
|
|
#define __emulate_2op_cl(ctxt, _op, _suffix, _type) \
|
2011-04-20 10:12:27 +00:00
|
|
|
do { \
|
|
|
|
|
unsigned long _tmp; \
|
2011-09-07 13:41:36 +00:00
|
|
|
_type _clv = (ctxt)->src2.val; \
|
|
|
|
|
_type _srcv = (ctxt)->src.val; \
|
|
|
|
|
_type _dstv = (ctxt)->dst.val; \
|
2011-04-20 10:12:27 +00:00
|
|
|
\
|
|
|
|
|
__asm__ __volatile__ ( \
|
|
|
|
|
_PRE_EFLAGS("0", "5", "2") \
|
|
|
|
|
_op _suffix " %4,%1 \n" \
|
|
|
|
|
_POST_EFLAGS("0", "5", "2") \
|
2011-09-07 13:41:36 +00:00
|
|
|
: "=m" ((ctxt)->eflags), "+r" (_dstv), "=&r" (_tmp) \
|
2011-04-20 10:12:27 +00:00
|
|
|
: "c" (_clv) , "r" (_srcv), "i" (EFLAGS_MASK) \
|
|
|
|
|
); \
|
|
|
|
|
\
|
2011-09-07 13:41:36 +00:00
|
|
|
(ctxt)->src2.val = (unsigned long) _clv; \
|
|
|
|
|
(ctxt)->src2.val = (unsigned long) _srcv; \
|
|
|
|
|
(ctxt)->dst.val = (unsigned long) _dstv; \
|
2008-12-04 13:29:00 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
2011-09-07 13:41:36 +00:00
|
|
|
#define emulate_2op_cl(ctxt, _op) \
|
2011-04-20 10:12:27 +00:00
|
|
|
do { \
|
2011-09-07 13:41:36 +00:00
|
|
|
switch ((ctxt)->dst.bytes) { \
|
2011-04-20 10:12:27 +00:00
|
|
|
case 2: \
|
2011-09-07 13:41:37 +00:00
|
|
|
__emulate_2op_cl(ctxt, _op, "w", u16); \
|
2011-04-20 10:12:27 +00:00
|
|
|
break; \
|
|
|
|
|
case 4: \
|
2011-09-07 13:41:37 +00:00
|
|
|
__emulate_2op_cl(ctxt, _op, "l", u32); \
|
2011-04-20 10:12:27 +00:00
|
|
|
break; \
|
|
|
|
|
case 8: \
|
2011-09-07 13:41:37 +00:00
|
|
|
ON64(__emulate_2op_cl(ctxt, _op, "q", ulong)); \
|
2011-04-20 10:12:27 +00:00
|
|
|
break; \
|
|
|
|
|
} \
|
2008-12-04 13:29:00 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
2011-09-07 13:41:38 +00:00
|
|
|
#define __emulate_1op(ctxt, _op, _suffix) \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
do { \
|
|
|
|
|
unsigned long _tmp; \
|
|
|
|
|
\
|
2008-11-26 13:14:10 +00:00
|
|
|
__asm__ __volatile__ ( \
|
|
|
|
|
_PRE_EFLAGS("0", "3", "2") \
|
|
|
|
|
_op _suffix " %1; " \
|
|
|
|
|
_POST_EFLAGS("0", "3", "2") \
|
2011-09-07 13:41:38 +00:00
|
|
|
: "=m" ((ctxt)->eflags), "+m" ((ctxt)->dst.val), \
|
2008-11-26 13:14:10 +00:00
|
|
|
"=&r" (_tmp) \
|
|
|
|
|
: "i" (EFLAGS_MASK)); \
|
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
|
|
/* Instruction has only one explicit operand (no source operand). */
|
2011-09-07 13:41:38 +00:00
|
|
|
#define emulate_1op(ctxt, _op) \
|
2008-11-26 13:14:10 +00:00
|
|
|
do { \
|
2011-09-07 13:41:38 +00:00
|
|
|
switch ((ctxt)->dst.bytes) { \
|
|
|
|
|
case 1: __emulate_1op(ctxt, _op, "b"); break; \
|
|
|
|
|
case 2: __emulate_1op(ctxt, _op, "w"); break; \
|
|
|
|
|
case 4: __emulate_1op(ctxt, _op, "l"); break; \
|
|
|
|
|
case 8: ON64(__emulate_1op(ctxt, _op, "q")); break; \
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
} \
|
|
|
|
|
} while (0)
|
|
|
|
|
|
2011-09-07 13:41:40 +00:00
|
|
|
#define __emulate_1op_rax_rdx(ctxt, _op, _suffix, _ex) \
|
2010-08-26 08:59:00 +00:00
|
|
|
do { \
|
|
|
|
|
unsigned long _tmp; \
|
2012-08-27 20:46:17 +00:00
|
|
|
ulong *rax = reg_rmw((ctxt), VCPU_REGS_RAX); \
|
|
|
|
|
ulong *rdx = reg_rmw((ctxt), VCPU_REGS_RDX); \
|
2010-08-26 08:59:00 +00:00
|
|
|
\
|
|
|
|
|
__asm__ __volatile__ ( \
|
|
|
|
|
_PRE_EFLAGS("0", "5", "1") \
|
|
|
|
|
"1: \n\t" \
|
|
|
|
|
_op _suffix " %6; " \
|
|
|
|
|
"2: \n\t" \
|
|
|
|
|
_POST_EFLAGS("0", "5", "1") \
|
|
|
|
|
".pushsection .fixup,\"ax\" \n\t" \
|
|
|
|
|
"3: movb $1, %4 \n\t" \
|
|
|
|
|
"jmp 2b \n\t" \
|
|
|
|
|
".popsection \n\t" \
|
|
|
|
|
_ASM_EXTABLE(1b, 3b) \
|
2011-09-07 13:41:40 +00:00
|
|
|
: "=m" ((ctxt)->eflags), "=&r" (_tmp), \
|
|
|
|
|
"+a" (*rax), "+d" (*rdx), "+qm"(_ex) \
|
|
|
|
|
: "i" (EFLAGS_MASK), "m" ((ctxt)->src.val), \
|
|
|
|
|
"a" (*rax), "d" (*rdx)); \
|
2010-08-26 08:59:00 +00:00
|
|
|
} while (0)
|
|
|
|
|
|
2010-08-08 18:11:37 +00:00
|
|
|
/* instruction has only one source operand, destination is implicit (e.g. mul, div, imul, idiv) */
|
2011-09-07 13:41:40 +00:00
|
|
|
#define emulate_1op_rax_rdx(ctxt, _op, _ex) \
|
2011-04-20 10:12:27 +00:00
|
|
|
do { \
|
2011-09-07 13:41:40 +00:00
|
|
|
switch((ctxt)->src.bytes) { \
|
2011-04-20 10:12:27 +00:00
|
|
|
case 1: \
|
2011-09-07 13:41:40 +00:00
|
|
|
__emulate_1op_rax_rdx(ctxt, _op, "b", _ex); \
|
2011-04-20 10:12:27 +00:00
|
|
|
break; \
|
|
|
|
|
case 2: \
|
2011-09-07 13:41:40 +00:00
|
|
|
__emulate_1op_rax_rdx(ctxt, _op, "w", _ex); \
|
2011-04-20 10:12:27 +00:00
|
|
|
break; \
|
|
|
|
|
case 4: \
|
2011-09-07 13:41:40 +00:00
|
|
|
__emulate_1op_rax_rdx(ctxt, _op, "l", _ex); \
|
2010-08-26 08:59:00 +00:00
|
|
|
break; \
|
|
|
|
|
case 8: ON64( \
|
2011-09-07 13:41:40 +00:00
|
|
|
__emulate_1op_rax_rdx(ctxt, _op, "q", _ex)); \
|
2010-08-26 08:59:00 +00:00
|
|
|
break; \
|
|
|
|
|
} \
|
|
|
|
|
} while (0)
|
|
|
|
|
|
2011-04-04 10:39:27 +00:00
|
|
|
static int emulator_check_intercept(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
enum x86_intercept intercept,
|
|
|
|
|
enum x86_intercept_stage stage)
|
|
|
|
|
{
|
|
|
|
|
struct x86_instruction_info info = {
|
|
|
|
|
.intercept = intercept,
|
2011-06-01 12:34:25 +00:00
|
|
|
.rep_prefix = ctxt->rep_prefix,
|
|
|
|
|
.modrm_mod = ctxt->modrm_mod,
|
|
|
|
|
.modrm_reg = ctxt->modrm_reg,
|
|
|
|
|
.modrm_rm = ctxt->modrm_rm,
|
|
|
|
|
.src_val = ctxt->src.val64,
|
|
|
|
|
.src_bytes = ctxt->src.bytes,
|
|
|
|
|
.dst_bytes = ctxt->dst.bytes,
|
|
|
|
|
.ad_bytes = ctxt->ad_bytes,
|
2011-04-04 10:39:27 +00:00
|
|
|
.next_rip = ctxt->eip,
|
|
|
|
|
};
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
return ctxt->ops->intercept(ctxt, &info, stage);
|
2011-04-04 10:39:27 +00:00
|
|
|
}
|
|
|
|
|
|
2012-06-07 14:49:24 +00:00
|
|
|
static void assign_masked(ulong *dest, ulong src, ulong mask)
|
|
|
|
|
{
|
|
|
|
|
*dest = (*dest & ~mask) | (src & mask);
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
static inline unsigned long ad_mask(struct x86_emulate_ctxt *ctxt)
|
2008-02-18 19:12:48 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
return (1UL << (ctxt->ad_bytes << 3)) - 1;
|
2008-02-18 19:12:48 +00:00
|
|
|
}
|
|
|
|
|
|
2012-06-07 14:49:24 +00:00
|
|
|
static ulong stack_mask(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u16 sel;
|
|
|
|
|
struct desc_struct ss;
|
|
|
|
|
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
return ~0UL;
|
|
|
|
|
ctxt->ops->get_segment(ctxt, &sel, &ss, NULL, VCPU_SREG_SS);
|
|
|
|
|
return ~0U >> ((ss.d ^ 1) * 16); /* d=0: 0xffff; d=1: 0xffffffff */
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-12 17:03:23 +00:00
|
|
|
static int stack_size(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
return (__fls(stack_mask(ctxt)) + 1) >> 3;
|
|
|
|
|
}
|
|
|
|
|
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
/* Access/update address held in a register, based on addressing mode. */
|
2008-02-19 15:40:38 +00:00
|
|
|
static inline unsigned long
|
2011-06-01 12:34:25 +00:00
|
|
|
address_mask(struct x86_emulate_ctxt *ctxt, unsigned long reg)
|
2008-02-19 15:40:38 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->ad_bytes == sizeof(unsigned long))
|
2008-02-19 15:40:38 +00:00
|
|
|
return reg;
|
|
|
|
|
else
|
2011-06-01 12:34:25 +00:00
|
|
|
return reg & ad_mask(ctxt);
|
2008-02-19 15:40:38 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline unsigned long
|
2011-06-01 12:34:25 +00:00
|
|
|
register_address(struct x86_emulate_ctxt *ctxt, unsigned long reg)
|
2008-02-19 15:40:38 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
return address_mask(ctxt, reg);
|
2008-02-19 15:40:38 +00:00
|
|
|
}
|
|
|
|
|
|
2012-08-19 11:34:31 +00:00
|
|
|
static void masked_increment(ulong *reg, ulong mask, int inc)
|
|
|
|
|
{
|
|
|
|
|
assign_masked(reg, *reg + inc, mask);
|
|
|
|
|
}
|
|
|
|
|
|
2008-02-19 15:40:41 +00:00
|
|
|
static inline void
|
2011-06-01 12:34:25 +00:00
|
|
|
register_address_increment(struct x86_emulate_ctxt *ctxt, unsigned long *reg, int inc)
|
2008-02-19 15:40:41 +00:00
|
|
|
{
|
2012-08-19 11:34:31 +00:00
|
|
|
ulong mask;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->ad_bytes == sizeof(unsigned long))
|
2012-08-19 11:34:31 +00:00
|
|
|
mask = ~0UL;
|
2008-02-19 15:40:41 +00:00
|
|
|
else
|
2012-08-19 11:34:31 +00:00
|
|
|
mask = ad_mask(ctxt);
|
|
|
|
|
masked_increment(reg, mask, inc);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void rsp_increment(struct x86_emulate_ctxt *ctxt, int inc)
|
|
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
masked_increment(reg_rmw(ctxt, VCPU_REGS_RSP), stack_mask(ctxt), inc);
|
2008-02-19 15:40:41 +00:00
|
|
|
}
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
static inline void jmp_rel(struct x86_emulate_ctxt *ctxt, int rel)
|
2008-02-19 15:40:41 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
register_address_increment(ctxt, &ctxt->_eip, rel);
|
2008-02-19 15:40:41 +00:00
|
|
|
}
|
2007-08-19 08:00:36 +00:00
|
|
|
|
2011-04-03 11:08:51 +00:00
|
|
|
static u32 desc_limit_scaled(struct desc_struct *desc)
|
|
|
|
|
{
|
|
|
|
|
u32 limit = get_desc_limit(desc);
|
|
|
|
|
|
|
|
|
|
return desc->g ? (limit << 12) | 0xfff : limit;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
static void set_seg_override(struct x86_emulate_ctxt *ctxt, int seg)
|
2008-06-22 13:22:51 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->has_seg_override = true;
|
|
|
|
|
ctxt->seg_override = seg;
|
2008-06-22 13:22:51 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
static unsigned long seg_base(struct x86_emulate_ctxt *ctxt, int seg)
|
2008-06-22 13:22:51 +00:00
|
|
|
{
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64 && seg < VCPU_SREG_FS)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
return ctxt->ops->get_cached_segment_base(ctxt, seg);
|
2008-06-22 13:22:51 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
static unsigned seg_override(struct x86_emulate_ctxt *ctxt)
|
2008-06-22 13:22:51 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
if (!ctxt->has_seg_override)
|
2008-06-22 13:22:51 +00:00
|
|
|
return 0;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
return ctxt->seg_override;
|
2008-06-22 13:22:51 +00:00
|
|
|
}
|
|
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
static int emulate_exception(struct x86_emulate_ctxt *ctxt, int vec,
|
|
|
|
|
u32 error, bool valid)
|
2010-04-28 16:15:44 +00:00
|
|
|
{
|
2010-11-22 15:53:21 +00:00
|
|
|
ctxt->exception.vector = vec;
|
|
|
|
|
ctxt->exception.error_code = error;
|
|
|
|
|
ctxt->exception.error_code_valid = valid;
|
2010-11-22 15:53:25 +00:00
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
2010-04-28 16:15:44 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:29 +00:00
|
|
|
static int emulate_db(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
return emulate_exception(ctxt, DB_VECTOR, 0, false);
|
|
|
|
|
}
|
|
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
static int emulate_gp(struct x86_emulate_ctxt *ctxt, int err)
|
2010-04-28 16:15:44 +00:00
|
|
|
{
|
2010-11-22 15:53:25 +00:00
|
|
|
return emulate_exception(ctxt, GP_VECTOR, err, true);
|
2010-04-28 16:15:44 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-03 09:32:09 +00:00
|
|
|
static int emulate_ss(struct x86_emulate_ctxt *ctxt, int err)
|
|
|
|
|
{
|
|
|
|
|
return emulate_exception(ctxt, SS_VECTOR, err, true);
|
|
|
|
|
}
|
|
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
static int emulate_ud(struct x86_emulate_ctxt *ctxt)
|
2010-04-28 16:15:44 +00:00
|
|
|
{
|
2010-11-22 15:53:25 +00:00
|
|
|
return emulate_exception(ctxt, UD_VECTOR, 0, false);
|
2010-04-28 16:15:44 +00:00
|
|
|
}
|
|
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
static int emulate_ts(struct x86_emulate_ctxt *ctxt, int err)
|
2010-04-28 16:15:44 +00:00
|
|
|
{
|
2010-11-22 15:53:25 +00:00
|
|
|
return emulate_exception(ctxt, TS_VECTOR, err, true);
|
2010-04-28 16:15:44 +00:00
|
|
|
}
|
|
|
|
|
|
2010-08-26 08:59:01 +00:00
|
|
|
static int emulate_de(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2010-11-22 15:53:25 +00:00
|
|
|
return emulate_exception(ctxt, DE_VECTOR, 0, false);
|
2010-08-26 08:59:01 +00:00
|
|
|
}
|
|
|
|
|
|
2011-03-29 09:41:27 +00:00
|
|
|
static int emulate_nm(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
return emulate_exception(ctxt, NM_VECTOR, 0, false);
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
static u16 get_segment_selector(struct x86_emulate_ctxt *ctxt, unsigned seg)
|
|
|
|
|
{
|
|
|
|
|
u16 selector;
|
|
|
|
|
struct desc_struct desc;
|
|
|
|
|
|
|
|
|
|
ctxt->ops->get_segment(ctxt, &selector, &desc, NULL, seg);
|
|
|
|
|
return selector;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void set_segment_selector(struct x86_emulate_ctxt *ctxt, u16 selector,
|
|
|
|
|
unsigned seg)
|
|
|
|
|
{
|
|
|
|
|
u16 dummy;
|
|
|
|
|
u32 base3;
|
|
|
|
|
struct desc_struct desc;
|
|
|
|
|
|
|
|
|
|
ctxt->ops->get_segment(ctxt, &dummy, &desc, &base3, seg);
|
|
|
|
|
ctxt->ops->set_segment(ctxt, selector, &desc, base3, seg);
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-09 15:39:59 +00:00
|
|
|
/*
|
|
|
|
|
* x86 defines three classes of vector instructions: explicitly
|
|
|
|
|
* aligned, explicitly unaligned, and the rest, which change behaviour
|
|
|
|
|
* depending on whether they're AVX encoded or not.
|
|
|
|
|
*
|
|
|
|
|
* Also included is CMPXCHG16B which is not a vector instruction, yet it is
|
|
|
|
|
* subject to the same check.
|
|
|
|
|
*/
|
|
|
|
|
static bool insn_aligned(struct x86_emulate_ctxt *ctxt, unsigned size)
|
|
|
|
|
{
|
|
|
|
|
if (likely(size < 16))
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
if (ctxt->d & Aligned)
|
|
|
|
|
return true;
|
|
|
|
|
else if (ctxt->d & Unaligned)
|
|
|
|
|
return false;
|
|
|
|
|
else if (ctxt->d & Avx)
|
|
|
|
|
return false;
|
|
|
|
|
else
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-18 16:05:53 +00:00
|
|
|
static int __linearize(struct x86_emulate_ctxt *ctxt,
|
2011-04-03 09:33:12 +00:00
|
|
|
struct segmented_address addr,
|
2011-04-18 16:05:53 +00:00
|
|
|
unsigned size, bool write, bool fetch,
|
2011-04-03 09:33:12 +00:00
|
|
|
ulong *linear)
|
|
|
|
|
{
|
2011-04-03 09:32:09 +00:00
|
|
|
struct desc_struct desc;
|
|
|
|
|
bool usable;
|
2011-04-03 09:33:12 +00:00
|
|
|
ulong la;
|
2011-04-03 09:32:09 +00:00
|
|
|
u32 lim;
|
2011-04-27 10:20:30 +00:00
|
|
|
u16 sel;
|
2011-04-03 09:32:09 +00:00
|
|
|
unsigned cpl, rpl;
|
2011-04-03 09:33:12 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
la = seg_base(ctxt, addr.seg) + addr.ea;
|
2011-04-03 09:32:09 +00:00
|
|
|
switch (ctxt->mode) {
|
|
|
|
|
case X86EMUL_MODE_REAL:
|
|
|
|
|
break;
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
|
|
|
|
if (((signed long)la << 16) >> 16 != la)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
break;
|
|
|
|
|
default:
|
2011-04-27 10:20:30 +00:00
|
|
|
usable = ctxt->ops->get_segment(ctxt, &sel, &desc, NULL,
|
|
|
|
|
addr.seg);
|
2011-04-03 09:32:09 +00:00
|
|
|
if (!usable)
|
|
|
|
|
goto bad;
|
|
|
|
|
/* code segment or read-only data segment */
|
|
|
|
|
if (((desc.type & 8) || !(desc.type & 2)) && write)
|
|
|
|
|
goto bad;
|
|
|
|
|
/* unreadable code segment */
|
2011-04-18 16:05:53 +00:00
|
|
|
if (!fetch && (desc.type & 8) && !(desc.type & 2))
|
2011-04-03 09:32:09 +00:00
|
|
|
goto bad;
|
|
|
|
|
lim = desc_limit_scaled(&desc);
|
|
|
|
|
if ((desc.type & 8) || !(desc.type & 4)) {
|
|
|
|
|
/* expand-up segment */
|
|
|
|
|
if (addr.ea > lim || (u32)(addr.ea + size - 1) > lim)
|
|
|
|
|
goto bad;
|
|
|
|
|
} else {
|
2012-06-28 07:19:51 +00:00
|
|
|
/* expand-down segment */
|
2011-04-03 09:32:09 +00:00
|
|
|
if (addr.ea <= lim || (u32)(addr.ea + size - 1) <= lim)
|
|
|
|
|
goto bad;
|
|
|
|
|
lim = desc.d ? 0xffffffff : 0xffff;
|
|
|
|
|
if (addr.ea > lim || (u32)(addr.ea + size - 1) > lim)
|
|
|
|
|
goto bad;
|
|
|
|
|
}
|
2011-04-20 10:37:53 +00:00
|
|
|
cpl = ctxt->ops->cpl(ctxt);
|
2011-04-27 10:20:30 +00:00
|
|
|
rpl = sel & 3;
|
2011-04-03 09:32:09 +00:00
|
|
|
cpl = max(cpl, rpl);
|
|
|
|
|
if (!(desc.type & 8)) {
|
|
|
|
|
/* data segment */
|
|
|
|
|
if (cpl > desc.dpl)
|
|
|
|
|
goto bad;
|
|
|
|
|
} else if ((desc.type & 8) && !(desc.type & 4)) {
|
|
|
|
|
/* nonconforming code segment */
|
|
|
|
|
if (cpl != desc.dpl)
|
|
|
|
|
goto bad;
|
|
|
|
|
} else if ((desc.type & 8) && (desc.type & 4)) {
|
|
|
|
|
/* conforming code segment */
|
|
|
|
|
if (cpl < desc.dpl)
|
|
|
|
|
goto bad;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
}
|
2011-06-01 12:34:25 +00:00
|
|
|
if (fetch ? ctxt->mode != X86EMUL_MODE_PROT64 : ctxt->ad_bytes != 8)
|
2011-04-03 09:33:12 +00:00
|
|
|
la &= (u32)-1;
|
2012-04-09 15:39:59 +00:00
|
|
|
if (insn_aligned(ctxt, size) && ((la & (size - 1)) != 0))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2011-04-03 09:33:12 +00:00
|
|
|
*linear = la;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
2011-04-03 09:32:09 +00:00
|
|
|
bad:
|
|
|
|
|
if (addr.seg == VCPU_SREG_SS)
|
|
|
|
|
return emulate_ss(ctxt, addr.seg);
|
|
|
|
|
else
|
|
|
|
|
return emulate_gp(ctxt, addr.seg);
|
2011-04-03 09:33:12 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-18 16:05:53 +00:00
|
|
|
static int linearize(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct segmented_address addr,
|
|
|
|
|
unsigned size, bool write,
|
|
|
|
|
ulong *linear)
|
|
|
|
|
{
|
|
|
|
|
return __linearize(ctxt, addr, size, write, false, linear);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2011-03-31 14:52:26 +00:00
|
|
|
static int segmented_read_std(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct segmented_address addr,
|
|
|
|
|
void *data,
|
|
|
|
|
unsigned size)
|
|
|
|
|
{
|
2011-03-31 16:54:30 +00:00
|
|
|
int rc;
|
|
|
|
|
ulong linear;
|
|
|
|
|
|
2011-04-03 08:31:19 +00:00
|
|
|
rc = linearize(ctxt, addr, size, false, &linear);
|
2011-03-31 16:54:30 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2011-04-20 10:37:53 +00:00
|
|
|
return ctxt->ops->read_std(ctxt, linear, data, size, &ctxt->exception);
|
2011-03-31 14:52:26 +00:00
|
|
|
}
|
|
|
|
|
|
2011-07-30 09:00:17 +00:00
|
|
|
/*
|
|
|
|
|
* Fetch the next byte of the instruction being emulated which is pointed to
|
|
|
|
|
* by ctxt->_eip, then increment ctxt->_eip.
|
|
|
|
|
*
|
|
|
|
|
* Also prefetch the remaining bytes of the instruction without crossing page
|
|
|
|
|
* boundary if they are not in fetch_cache yet.
|
|
|
|
|
*/
|
|
|
|
|
static int do_insn_fetch_byte(struct x86_emulate_ctxt *ctxt, u8 *dest)
|
2007-11-20 11:15:52 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
struct fetch_cache *fc = &ctxt->fetch;
|
2007-11-20 11:15:52 +00:00
|
|
|
int rc;
|
2010-04-11 10:05:15 +00:00
|
|
|
int size, cur_size;
|
2007-11-20 11:15:52 +00:00
|
|
|
|
2011-07-30 09:00:17 +00:00
|
|
|
if (ctxt->_eip == fc->end) {
|
2011-04-18 16:05:53 +00:00
|
|
|
unsigned long linear;
|
2011-07-30 09:00:17 +00:00
|
|
|
struct segmented_address addr = { .seg = VCPU_SREG_CS,
|
|
|
|
|
.ea = ctxt->_eip };
|
2010-04-11 10:05:15 +00:00
|
|
|
cur_size = fc->end - fc->start;
|
2011-07-30 09:00:17 +00:00
|
|
|
size = min(15UL - cur_size,
|
|
|
|
|
PAGE_SIZE - offset_in_page(ctxt->_eip));
|
2011-04-18 16:05:53 +00:00
|
|
|
rc = __linearize(ctxt, addr, size, false, true, &linear);
|
2011-07-30 09:02:29 +00:00
|
|
|
if (unlikely(rc != X86EMUL_CONTINUE))
|
2011-04-18 16:05:53 +00:00
|
|
|
return rc;
|
2011-05-14 15:57:43 +00:00
|
|
|
rc = ctxt->ops->fetch(ctxt, linear, fc->data + cur_size,
|
|
|
|
|
size, &ctxt->exception);
|
2011-07-30 09:02:29 +00:00
|
|
|
if (unlikely(rc != X86EMUL_CONTINUE))
|
2007-11-20 11:15:52 +00:00
|
|
|
return rc;
|
2010-04-11 10:05:15 +00:00
|
|
|
fc->end += size;
|
2007-11-20 11:15:52 +00:00
|
|
|
}
|
2011-07-30 09:00:17 +00:00
|
|
|
*dest = fc->data[ctxt->_eip - fc->start];
|
|
|
|
|
ctxt->_eip++;
|
2010-02-12 06:53:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2007-11-20 11:15:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int do_insn_fetch(struct x86_emulate_ctxt *ctxt,
|
2011-07-30 09:00:17 +00:00
|
|
|
void *dest, unsigned size)
|
2007-11-20 11:15:52 +00:00
|
|
|
{
|
2010-02-12 06:53:59 +00:00
|
|
|
int rc;
|
2007-11-20 11:15:52 +00:00
|
|
|
|
2009-11-24 13:20:15 +00:00
|
|
|
/* x86 instructions are limited to 15 bytes. */
|
2011-07-30 09:02:29 +00:00
|
|
|
if (unlikely(ctxt->_eip + size - ctxt->eip > 15))
|
2009-11-24 13:20:15 +00:00
|
|
|
return X86EMUL_UNHANDLEABLE;
|
2007-11-20 11:15:52 +00:00
|
|
|
while (size--) {
|
2011-07-30 09:00:17 +00:00
|
|
|
rc = do_insn_fetch_byte(ctxt, dest++);
|
2010-02-12 06:53:59 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
2007-11-20 11:15:52 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
2010-02-12 06:53:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2007-11-20 11:15:52 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-14 15:54:58 +00:00
|
|
|
/* Fetch next part of the instruction being emulated. */
|
2011-07-30 09:01:26 +00:00
|
|
|
#define insn_fetch(_type, _ctxt) \
|
2011-05-14 15:54:58 +00:00
|
|
|
({ unsigned long _x; \
|
2011-07-30 09:01:26 +00:00
|
|
|
rc = do_insn_fetch(_ctxt, &_x, sizeof(_type)); \
|
2011-05-14 15:54:58 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE) \
|
|
|
|
|
goto done; \
|
|
|
|
|
(_type)_x; \
|
|
|
|
|
})
|
|
|
|
|
|
2011-07-30 09:00:17 +00:00
|
|
|
#define insn_fetch_arr(_arr, _size, _ctxt) \
|
|
|
|
|
({ rc = do_insn_fetch(_ctxt, _arr, (_size)); \
|
2011-05-14 15:54:58 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE) \
|
|
|
|
|
goto done; \
|
|
|
|
|
})
|
|
|
|
|
|
2007-07-17 13:16:11 +00:00
|
|
|
/*
|
|
|
|
|
* Given the 'reg' portion of a ModRM byte, and a register block, return a
|
|
|
|
|
* pointer into the block that addresses the relevant register.
|
|
|
|
|
* @highbyte_regs specifies whether to decode AH,CH,DH,BH.
|
|
|
|
|
*/
|
2012-08-27 20:46:17 +00:00
|
|
|
static void *decode_register(struct x86_emulate_ctxt *ctxt, u8 modrm_reg,
|
2007-07-17 13:16:11 +00:00
|
|
|
int highbyte_regs)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
{
|
|
|
|
|
void *p;
|
|
|
|
|
|
|
|
|
|
if (highbyte_regs && modrm_reg >= 4 && modrm_reg < 8)
|
2012-08-27 20:46:17 +00:00
|
|
|
p = (unsigned char *)reg_rmw(ctxt, modrm_reg & 3) + 1;
|
|
|
|
|
else
|
|
|
|
|
p = reg_rmw(ctxt, modrm_reg);
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
return p;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int read_descriptor(struct x86_emulate_ctxt *ctxt,
|
2010-11-17 13:28:21 +00:00
|
|
|
struct segmented_address addr,
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
u16 *size, unsigned long *address, int op_bytes)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
if (op_bytes == 2)
|
|
|
|
|
op_bytes = 3;
|
|
|
|
|
*address = 0;
|
2011-03-31 14:52:26 +00:00
|
|
|
rc = segmented_read_std(ctxt, addr, size, 2);
|
2010-02-12 06:57:56 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
return rc;
|
2010-11-17 13:28:22 +00:00
|
|
|
addr.ea += 2;
|
2011-03-31 14:52:26 +00:00
|
|
|
rc = segmented_read_std(ctxt, addr, address, op_bytes);
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2007-09-15 07:23:07 +00:00
|
|
|
static int test_cc(unsigned int condition, unsigned int flags)
|
|
|
|
|
{
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
|
|
switch ((condition & 15) >> 1) {
|
|
|
|
|
case 0: /* o */
|
|
|
|
|
rc |= (flags & EFLG_OF);
|
|
|
|
|
break;
|
|
|
|
|
case 1: /* b/c/nae */
|
|
|
|
|
rc |= (flags & EFLG_CF);
|
|
|
|
|
break;
|
|
|
|
|
case 2: /* z/e */
|
|
|
|
|
rc |= (flags & EFLG_ZF);
|
|
|
|
|
break;
|
|
|
|
|
case 3: /* be/na */
|
|
|
|
|
rc |= (flags & (EFLG_CF|EFLG_ZF));
|
|
|
|
|
break;
|
|
|
|
|
case 4: /* s */
|
|
|
|
|
rc |= (flags & EFLG_SF);
|
|
|
|
|
break;
|
|
|
|
|
case 5: /* p/pe */
|
|
|
|
|
rc |= (flags & EFLG_PF);
|
|
|
|
|
break;
|
|
|
|
|
case 7: /* le/ng */
|
|
|
|
|
rc |= (flags & EFLG_ZF);
|
|
|
|
|
/* fall through */
|
|
|
|
|
case 6: /* l/nge */
|
|
|
|
|
rc |= (!(flags & EFLG_SF) != !(flags & EFLG_OF));
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Odd condition identifiers (lsb == 1) have inverted sense. */
|
|
|
|
|
return (!!rc ^ (condition & 1));
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-01 09:53:09 +00:00
|
|
|
static void fetch_register_operand(struct operand *op)
|
|
|
|
|
{
|
|
|
|
|
switch (op->bytes) {
|
|
|
|
|
case 1:
|
|
|
|
|
op->val = *(u8 *)op->addr.reg;
|
|
|
|
|
break;
|
|
|
|
|
case 2:
|
|
|
|
|
op->val = *(u16 *)op->addr.reg;
|
|
|
|
|
break;
|
|
|
|
|
case 4:
|
|
|
|
|
op->val = *(u32 *)op->addr.reg;
|
|
|
|
|
break;
|
|
|
|
|
case 8:
|
|
|
|
|
op->val = *(u64 *)op->addr.reg;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-03-29 09:41:27 +00:00
|
|
|
static void read_sse_reg(struct x86_emulate_ctxt *ctxt, sse128_t *data, int reg)
|
|
|
|
|
{
|
|
|
|
|
ctxt->ops->get_fpu(ctxt);
|
|
|
|
|
switch (reg) {
|
|
|
|
|
case 0: asm("movdqu %%xmm0, %0" : "=m"(*data)); break;
|
|
|
|
|
case 1: asm("movdqu %%xmm1, %0" : "=m"(*data)); break;
|
|
|
|
|
case 2: asm("movdqu %%xmm2, %0" : "=m"(*data)); break;
|
|
|
|
|
case 3: asm("movdqu %%xmm3, %0" : "=m"(*data)); break;
|
|
|
|
|
case 4: asm("movdqu %%xmm4, %0" : "=m"(*data)); break;
|
|
|
|
|
case 5: asm("movdqu %%xmm5, %0" : "=m"(*data)); break;
|
|
|
|
|
case 6: asm("movdqu %%xmm6, %0" : "=m"(*data)); break;
|
|
|
|
|
case 7: asm("movdqu %%xmm7, %0" : "=m"(*data)); break;
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
case 8: asm("movdqu %%xmm8, %0" : "=m"(*data)); break;
|
|
|
|
|
case 9: asm("movdqu %%xmm9, %0" : "=m"(*data)); break;
|
|
|
|
|
case 10: asm("movdqu %%xmm10, %0" : "=m"(*data)); break;
|
|
|
|
|
case 11: asm("movdqu %%xmm11, %0" : "=m"(*data)); break;
|
|
|
|
|
case 12: asm("movdqu %%xmm12, %0" : "=m"(*data)); break;
|
|
|
|
|
case 13: asm("movdqu %%xmm13, %0" : "=m"(*data)); break;
|
|
|
|
|
case 14: asm("movdqu %%xmm14, %0" : "=m"(*data)); break;
|
|
|
|
|
case 15: asm("movdqu %%xmm15, %0" : "=m"(*data)); break;
|
|
|
|
|
#endif
|
|
|
|
|
default: BUG();
|
|
|
|
|
}
|
|
|
|
|
ctxt->ops->put_fpu(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void write_sse_reg(struct x86_emulate_ctxt *ctxt, sse128_t *data,
|
|
|
|
|
int reg)
|
|
|
|
|
{
|
|
|
|
|
ctxt->ops->get_fpu(ctxt);
|
|
|
|
|
switch (reg) {
|
|
|
|
|
case 0: asm("movdqu %0, %%xmm0" : : "m"(*data)); break;
|
|
|
|
|
case 1: asm("movdqu %0, %%xmm1" : : "m"(*data)); break;
|
|
|
|
|
case 2: asm("movdqu %0, %%xmm2" : : "m"(*data)); break;
|
|
|
|
|
case 3: asm("movdqu %0, %%xmm3" : : "m"(*data)); break;
|
|
|
|
|
case 4: asm("movdqu %0, %%xmm4" : : "m"(*data)); break;
|
|
|
|
|
case 5: asm("movdqu %0, %%xmm5" : : "m"(*data)); break;
|
|
|
|
|
case 6: asm("movdqu %0, %%xmm6" : : "m"(*data)); break;
|
|
|
|
|
case 7: asm("movdqu %0, %%xmm7" : : "m"(*data)); break;
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
case 8: asm("movdqu %0, %%xmm8" : : "m"(*data)); break;
|
|
|
|
|
case 9: asm("movdqu %0, %%xmm9" : : "m"(*data)); break;
|
|
|
|
|
case 10: asm("movdqu %0, %%xmm10" : : "m"(*data)); break;
|
|
|
|
|
case 11: asm("movdqu %0, %%xmm11" : : "m"(*data)); break;
|
|
|
|
|
case 12: asm("movdqu %0, %%xmm12" : : "m"(*data)); break;
|
|
|
|
|
case 13: asm("movdqu %0, %%xmm13" : : "m"(*data)); break;
|
|
|
|
|
case 14: asm("movdqu %0, %%xmm14" : : "m"(*data)); break;
|
|
|
|
|
case 15: asm("movdqu %0, %%xmm15" : : "m"(*data)); break;
|
|
|
|
|
#endif
|
|
|
|
|
default: BUG();
|
|
|
|
|
}
|
|
|
|
|
ctxt->ops->put_fpu(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-09 15:40:02 +00:00
|
|
|
static void read_mmx_reg(struct x86_emulate_ctxt *ctxt, u64 *data, int reg)
|
|
|
|
|
{
|
|
|
|
|
ctxt->ops->get_fpu(ctxt);
|
|
|
|
|
switch (reg) {
|
|
|
|
|
case 0: asm("movq %%mm0, %0" : "=m"(*data)); break;
|
|
|
|
|
case 1: asm("movq %%mm1, %0" : "=m"(*data)); break;
|
|
|
|
|
case 2: asm("movq %%mm2, %0" : "=m"(*data)); break;
|
|
|
|
|
case 3: asm("movq %%mm3, %0" : "=m"(*data)); break;
|
|
|
|
|
case 4: asm("movq %%mm4, %0" : "=m"(*data)); break;
|
|
|
|
|
case 5: asm("movq %%mm5, %0" : "=m"(*data)); break;
|
|
|
|
|
case 6: asm("movq %%mm6, %0" : "=m"(*data)); break;
|
|
|
|
|
case 7: asm("movq %%mm7, %0" : "=m"(*data)); break;
|
|
|
|
|
default: BUG();
|
|
|
|
|
}
|
|
|
|
|
ctxt->ops->put_fpu(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void write_mmx_reg(struct x86_emulate_ctxt *ctxt, u64 *data, int reg)
|
|
|
|
|
{
|
|
|
|
|
ctxt->ops->get_fpu(ctxt);
|
|
|
|
|
switch (reg) {
|
|
|
|
|
case 0: asm("movq %0, %%mm0" : : "m"(*data)); break;
|
|
|
|
|
case 1: asm("movq %0, %%mm1" : : "m"(*data)); break;
|
|
|
|
|
case 2: asm("movq %0, %%mm2" : : "m"(*data)); break;
|
|
|
|
|
case 3: asm("movq %0, %%mm3" : : "m"(*data)); break;
|
|
|
|
|
case 4: asm("movq %0, %%mm4" : : "m"(*data)); break;
|
|
|
|
|
case 5: asm("movq %0, %%mm5" : : "m"(*data)); break;
|
|
|
|
|
case 6: asm("movq %0, %%mm6" : : "m"(*data)); break;
|
|
|
|
|
case 7: asm("movq %0, %%mm7" : : "m"(*data)); break;
|
|
|
|
|
default: BUG();
|
|
|
|
|
}
|
|
|
|
|
ctxt->ops->put_fpu(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2011-03-29 09:41:27 +00:00
|
|
|
static void decode_register_operand(struct x86_emulate_ctxt *ctxt,
|
2012-01-16 13:08:45 +00:00
|
|
|
struct operand *op)
|
2007-10-31 08:27:04 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
unsigned reg = ctxt->modrm_reg;
|
|
|
|
|
int highbyte_regs = ctxt->rex_prefix == 0;
|
2007-10-31 09:15:56 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (!(ctxt->d & ModRM))
|
|
|
|
|
reg = (ctxt->b & 7) | ((ctxt->rex_prefix & 1) << 3);
|
2011-03-29 09:41:27 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->d & Sse) {
|
2011-03-29 09:41:27 +00:00
|
|
|
op->type = OP_XMM;
|
|
|
|
|
op->bytes = 16;
|
|
|
|
|
op->addr.xmm = reg;
|
|
|
|
|
read_sse_reg(ctxt, &op->vec_val, reg);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2012-04-09 15:40:02 +00:00
|
|
|
if (ctxt->d & Mmx) {
|
|
|
|
|
reg &= 7;
|
|
|
|
|
op->type = OP_MM;
|
|
|
|
|
op->bytes = 8;
|
|
|
|
|
op->addr.mm = reg;
|
|
|
|
|
return;
|
|
|
|
|
}
|
2011-03-29 09:41:27 +00:00
|
|
|
|
2007-10-31 08:27:04 +00:00
|
|
|
op->type = OP_REG;
|
2012-01-16 13:08:45 +00:00
|
|
|
if (ctxt->d & ByteOp) {
|
2012-08-27 20:46:17 +00:00
|
|
|
op->addr.reg = decode_register(ctxt, reg, highbyte_regs);
|
2007-10-31 08:27:04 +00:00
|
|
|
op->bytes = 1;
|
|
|
|
|
} else {
|
2012-08-27 20:46:17 +00:00
|
|
|
op->addr.reg = decode_register(ctxt, reg, 0);
|
2011-06-01 12:34:25 +00:00
|
|
|
op->bytes = ctxt->op_bytes;
|
2007-10-31 08:27:04 +00:00
|
|
|
}
|
2010-08-01 09:53:09 +00:00
|
|
|
fetch_register_operand(op);
|
2007-10-31 08:27:04 +00:00
|
|
|
op->orig_val = op->val;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-10 14:15:39 +00:00
|
|
|
static void adjust_modrm_seg(struct x86_emulate_ctxt *ctxt, int base_reg)
|
|
|
|
|
{
|
|
|
|
|
if (base_reg == VCPU_REGS_RSP || base_reg == VCPU_REGS_RBP)
|
|
|
|
|
ctxt->modrm_seg = VCPU_SREG_SS;
|
|
|
|
|
}
|
|
|
|
|
|
2007-11-01 04:31:28 +00:00
|
|
|
static int decode_modrm(struct x86_emulate_ctxt *ctxt,
|
2010-08-01 12:40:19 +00:00
|
|
|
struct operand *op)
|
2007-11-01 04:31:28 +00:00
|
|
|
{
|
|
|
|
|
u8 sib;
|
2008-06-16 05:09:11 +00:00
|
|
|
int index_reg = 0, base_reg = 0, scale;
|
2010-02-12 06:53:59 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
2010-08-01 12:40:19 +00:00
|
|
|
ulong modrm_ea = 0;
|
2007-11-01 04:31:28 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->rex_prefix) {
|
|
|
|
|
ctxt->modrm_reg = (ctxt->rex_prefix & 4) << 1; /* REX.R */
|
|
|
|
|
index_reg = (ctxt->rex_prefix & 2) << 2; /* REX.X */
|
|
|
|
|
ctxt->modrm_rm = base_reg = (ctxt->rex_prefix & 1) << 3; /* REG.B */
|
2007-11-01 04:31:28 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->modrm_mod |= (ctxt->modrm & 0xc0) >> 6;
|
|
|
|
|
ctxt->modrm_reg |= (ctxt->modrm & 0x38) >> 3;
|
|
|
|
|
ctxt->modrm_rm |= (ctxt->modrm & 0x07);
|
|
|
|
|
ctxt->modrm_seg = VCPU_SREG_DS;
|
2007-11-01 04:31:28 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_mod == 3) {
|
2010-08-01 12:40:19 +00:00
|
|
|
op->type = OP_REG;
|
2011-06-01 12:34:25 +00:00
|
|
|
op->bytes = (ctxt->d & ByteOp) ? 1 : ctxt->op_bytes;
|
2012-08-27 20:46:17 +00:00
|
|
|
op->addr.reg = decode_register(ctxt, ctxt->modrm_rm, ctxt->d & ByteOp);
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->d & Sse) {
|
2011-03-29 09:41:27 +00:00
|
|
|
op->type = OP_XMM;
|
|
|
|
|
op->bytes = 16;
|
2011-06-01 12:34:25 +00:00
|
|
|
op->addr.xmm = ctxt->modrm_rm;
|
|
|
|
|
read_sse_reg(ctxt, &op->vec_val, ctxt->modrm_rm);
|
2011-03-29 09:41:27 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
2012-04-09 15:40:02 +00:00
|
|
|
if (ctxt->d & Mmx) {
|
|
|
|
|
op->type = OP_MM;
|
|
|
|
|
op->bytes = 8;
|
|
|
|
|
op->addr.xmm = ctxt->modrm_rm & 7;
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
2010-08-01 12:40:19 +00:00
|
|
|
fetch_register_operand(op);
|
2007-11-01 04:31:28 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-01 12:40:19 +00:00
|
|
|
op->type = OP_MEM;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->ad_bytes == 2) {
|
2012-08-27 20:46:17 +00:00
|
|
|
unsigned bx = reg_read(ctxt, VCPU_REGS_RBX);
|
|
|
|
|
unsigned bp = reg_read(ctxt, VCPU_REGS_RBP);
|
|
|
|
|
unsigned si = reg_read(ctxt, VCPU_REGS_RSI);
|
|
|
|
|
unsigned di = reg_read(ctxt, VCPU_REGS_RDI);
|
2007-11-01 04:31:28 +00:00
|
|
|
|
|
|
|
|
/* 16-bit ModR/M decode. */
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->modrm_mod) {
|
2007-11-01 04:31:28 +00:00
|
|
|
case 0:
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_rm == 6)
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(u16, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 1:
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(s8, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 2:
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(u16, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->modrm_rm) {
|
2007-11-01 04:31:28 +00:00
|
|
|
case 0:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += bx + si;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 1:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += bx + di;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 2:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += bp + si;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 3:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += bp + di;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 4:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += si;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 5:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += di;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 6:
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_mod != 0)
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += bp;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 7:
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea += bx;
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_rm == 2 || ctxt->modrm_rm == 3 ||
|
|
|
|
|
(ctxt->modrm_rm == 6 && ctxt->modrm_mod != 0))
|
|
|
|
|
ctxt->modrm_seg = VCPU_SREG_SS;
|
2010-08-01 12:40:19 +00:00
|
|
|
modrm_ea = (u16)modrm_ea;
|
2007-11-01 04:31:28 +00:00
|
|
|
} else {
|
|
|
|
|
/* 32/64-bit ModR/M decode. */
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->modrm_rm & 7) == 4) {
|
2011-07-30 09:01:26 +00:00
|
|
|
sib = insn_fetch(u8, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
index_reg |= (sib >> 3) & 7;
|
|
|
|
|
base_reg |= sib & 7;
|
|
|
|
|
scale = sib >> 6;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((base_reg & 7) == 5 && ctxt->modrm_mod == 0)
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(s32, ctxt);
|
2012-06-10 14:15:39 +00:00
|
|
|
else {
|
2012-08-27 20:46:17 +00:00
|
|
|
modrm_ea += reg_read(ctxt, base_reg);
|
2012-06-10 14:15:39 +00:00
|
|
|
adjust_modrm_seg(ctxt, base_reg);
|
|
|
|
|
}
|
2008-06-16 04:23:17 +00:00
|
|
|
if (index_reg != 4)
|
2012-08-27 20:46:17 +00:00
|
|
|
modrm_ea += reg_read(ctxt, index_reg) << scale;
|
2011-06-01 12:34:25 +00:00
|
|
|
} else if ((ctxt->modrm_rm & 7) == 5 && ctxt->modrm_mod == 0) {
|
2008-06-16 04:53:26 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->rip_relative = 1;
|
2012-06-10 14:15:39 +00:00
|
|
|
} else {
|
|
|
|
|
base_reg = ctxt->modrm_rm;
|
2012-08-27 20:46:17 +00:00
|
|
|
modrm_ea += reg_read(ctxt, base_reg);
|
2012-06-10 14:15:39 +00:00
|
|
|
adjust_modrm_seg(ctxt, base_reg);
|
|
|
|
|
}
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->modrm_mod) {
|
2007-11-01 04:31:28 +00:00
|
|
|
case 0:
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_rm == 5)
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(s32, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 1:
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(s8, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 2:
|
2011-07-30 09:01:26 +00:00
|
|
|
modrm_ea += insn_fetch(s32, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
2010-11-17 13:28:21 +00:00
|
|
|
op->addr.mem.ea = modrm_ea;
|
2007-11-01 04:31:28 +00:00
|
|
|
done:
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int decode_abs(struct x86_emulate_ctxt *ctxt,
|
2010-08-01 12:40:19 +00:00
|
|
|
struct operand *op)
|
2007-11-01 04:31:28 +00:00
|
|
|
{
|
2010-02-12 06:53:59 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
2007-11-01 04:31:28 +00:00
|
|
|
|
2010-08-01 12:40:19 +00:00
|
|
|
op->type = OP_MEM;
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->ad_bytes) {
|
2007-11-01 04:31:28 +00:00
|
|
|
case 2:
|
2011-07-30 09:01:26 +00:00
|
|
|
op->addr.mem.ea = insn_fetch(u16, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 4:
|
2011-07-30 09:01:26 +00:00
|
|
|
op->addr.mem.ea = insn_fetch(u32, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
case 8:
|
2011-07-30 09:01:26 +00:00
|
|
|
op->addr.mem.ea = insn_fetch(u64, ctxt);
|
2007-11-01 04:31:28 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
done:
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
static void fetch_bit_operand(struct x86_emulate_ctxt *ctxt)
|
2010-08-09 03:34:56 +00:00
|
|
|
{
|
2010-09-28 08:33:32 +00:00
|
|
|
long sv = 0, mask;
|
2010-08-09 03:34:56 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->dst.type == OP_MEM && ctxt->src.type == OP_REG) {
|
|
|
|
|
mask = ~(ctxt->dst.bytes * 8 - 1);
|
2010-08-09 03:34:56 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->src.bytes == 2)
|
|
|
|
|
sv = (s16)ctxt->src.val & (s16)mask;
|
|
|
|
|
else if (ctxt->src.bytes == 4)
|
|
|
|
|
sv = (s32)ctxt->src.val & (s32)mask;
|
2010-08-09 03:34:56 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.addr.mem.ea += (sv >> 3);
|
2010-08-09 03:34:56 +00:00
|
|
|
}
|
2010-08-09 03:39:14 +00:00
|
|
|
|
|
|
|
|
/* only subword offset */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val &= (ctxt->dst.bytes << 3) - 1;
|
2010-08-09 03:34:56 +00:00
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
static int read_emulated(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
unsigned long addr, void *dest, unsigned size)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
{
|
2010-07-29 12:11:52 +00:00
|
|
|
int rc;
|
2011-06-01 12:34:25 +00:00
|
|
|
struct read_cache *mc = &ctxt->mem_read;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2012-07-26 05:12:22 +00:00
|
|
|
if (mc->pos < mc->end)
|
|
|
|
|
goto read_cached;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2012-07-26 05:12:22 +00:00
|
|
|
WARN_ON((mc->end + size) >= sizeof(mc->data));
|
|
|
|
|
|
|
|
|
|
rc = ctxt->ops->read_emulated(ctxt, addr, mc->data + mc->end, size,
|
|
|
|
|
&ctxt->exception);
|
|
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
|
|
|
|
mc->end += size;
|
|
|
|
|
|
|
|
|
|
read_cached:
|
|
|
|
|
memcpy(dest, mc->data + mc->pos, size);
|
|
|
|
|
mc->pos += size;
|
2010-07-29 12:11:52 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2011-03-31 14:52:26 +00:00
|
|
|
static int segmented_read(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct segmented_address addr,
|
|
|
|
|
void *data,
|
|
|
|
|
unsigned size)
|
|
|
|
|
{
|
2011-03-31 16:54:30 +00:00
|
|
|
int rc;
|
|
|
|
|
ulong linear;
|
|
|
|
|
|
2011-04-03 08:31:19 +00:00
|
|
|
rc = linearize(ctxt, addr, size, false, &linear);
|
2011-03-31 16:54:30 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2011-05-14 16:00:52 +00:00
|
|
|
return read_emulated(ctxt, linear, data, size);
|
2011-03-31 14:52:26 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int segmented_write(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct segmented_address addr,
|
|
|
|
|
const void *data,
|
|
|
|
|
unsigned size)
|
|
|
|
|
{
|
2011-03-31 16:54:30 +00:00
|
|
|
int rc;
|
|
|
|
|
ulong linear;
|
|
|
|
|
|
2011-04-03 08:31:19 +00:00
|
|
|
rc = linearize(ctxt, addr, size, true, &linear);
|
2011-03-31 16:54:30 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2011-04-20 10:37:53 +00:00
|
|
|
return ctxt->ops->write_emulated(ctxt, linear, data, size,
|
|
|
|
|
&ctxt->exception);
|
2011-03-31 14:52:26 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int segmented_cmpxchg(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct segmented_address addr,
|
|
|
|
|
const void *orig_data, const void *data,
|
|
|
|
|
unsigned size)
|
|
|
|
|
{
|
2011-03-31 16:54:30 +00:00
|
|
|
int rc;
|
|
|
|
|
ulong linear;
|
|
|
|
|
|
2011-04-03 08:31:19 +00:00
|
|
|
rc = linearize(ctxt, addr, size, true, &linear);
|
2011-03-31 16:54:30 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2011-04-20 10:37:53 +00:00
|
|
|
return ctxt->ops->cmpxchg_emulated(ctxt, linear, orig_data, data,
|
|
|
|
|
size, &ctxt->exception);
|
2011-03-31 14:52:26 +00:00
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
static int pio_in_emulated(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
unsigned int size, unsigned short port,
|
|
|
|
|
void *dest)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
struct read_cache *rc = &ctxt->io_read;
|
2007-09-25 11:36:40 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc->pos == rc->end) { /* refill pio read ahead */
|
|
|
|
|
unsigned int in_page, n;
|
2011-06-01 12:34:25 +00:00
|
|
|
unsigned int count = ctxt->rep_prefix ?
|
2012-08-27 20:46:17 +00:00
|
|
|
address_mask(ctxt, reg_read(ctxt, VCPU_REGS_RCX)) : 1;
|
2010-07-29 12:11:52 +00:00
|
|
|
in_page = (ctxt->eflags & EFLG_DF) ?
|
2012-08-27 20:46:17 +00:00
|
|
|
offset_in_page(reg_read(ctxt, VCPU_REGS_RDI)) :
|
|
|
|
|
PAGE_SIZE - offset_in_page(reg_read(ctxt, VCPU_REGS_RDI));
|
2010-07-29 12:11:52 +00:00
|
|
|
n = min(min(in_page, (unsigned int)sizeof(rc->data)) / size,
|
|
|
|
|
count);
|
|
|
|
|
if (n == 0)
|
|
|
|
|
n = 1;
|
|
|
|
|
rc->pos = rc->end = 0;
|
2011-05-14 16:00:52 +00:00
|
|
|
if (!ctxt->ops->pio_in_emulated(ctxt, size, port, rc->data, n))
|
2010-07-29 12:11:52 +00:00
|
|
|
return 0;
|
|
|
|
|
rc->end = n * size;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
memcpy(dest, rc->data + rc->pos, size);
|
|
|
|
|
rc->pos += size;
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2012-02-08 13:34:38 +00:00
|
|
|
static int read_interrupt_descriptor(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 index, struct desc_struct *desc)
|
|
|
|
|
{
|
|
|
|
|
struct desc_ptr dt;
|
|
|
|
|
ulong addr;
|
|
|
|
|
|
|
|
|
|
ctxt->ops->get_idt(ctxt, &dt);
|
|
|
|
|
|
|
|
|
|
if (dt.size < index * 8 + 7)
|
|
|
|
|
return emulate_gp(ctxt, index << 3 | 0x2);
|
|
|
|
|
|
|
|
|
|
addr = dt.address + index * 8;
|
|
|
|
|
return ctxt->ops->read_std(ctxt, addr, desc, sizeof *desc,
|
|
|
|
|
&ctxt->exception);
|
|
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
static void get_descriptor_table_ptr(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 selector, struct desc_ptr *dt)
|
|
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if (selector & 1 << 2) {
|
|
|
|
|
struct desc_struct desc;
|
2011-04-27 10:20:30 +00:00
|
|
|
u16 sel;
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
memset (dt, 0, sizeof *dt);
|
2011-04-27 10:20:30 +00:00
|
|
|
if (!ops->get_segment(ctxt, &sel, &desc, NULL, VCPU_SREG_LDTR))
|
2010-07-29 12:11:52 +00:00
|
|
|
return;
|
2008-01-18 10:38:59 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
dt->size = desc_limit_scaled(&desc); /* what if limit > 65535? */
|
|
|
|
|
dt->address = get_desc_base(&desc);
|
|
|
|
|
} else
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_gdt(ctxt, dt);
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
2010-07-29 12:11:39 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
/* allowed just for 8 bytes segments */
|
|
|
|
|
static int read_segment_descriptor(struct x86_emulate_ctxt *ctxt,
|
2012-06-13 13:29:39 +00:00
|
|
|
u16 selector, struct desc_struct *desc,
|
|
|
|
|
ulong *desc_addr_p)
|
2010-07-29 12:11:52 +00:00
|
|
|
{
|
|
|
|
|
struct desc_ptr dt;
|
|
|
|
|
u16 index = selector >> 3;
|
|
|
|
|
ulong addr;
|
2010-07-29 12:11:39 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
get_descriptor_table_ptr(ctxt, selector, &dt);
|
2010-07-29 12:11:39 +00:00
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
if (dt.size < index * 8 + 7)
|
|
|
|
|
return emulate_gp(ctxt, selector & 0xfffc);
|
2008-01-18 10:38:59 +00:00
|
|
|
|
2012-06-13 13:29:39 +00:00
|
|
|
*desc_addr_p = addr = dt.address + index * 8;
|
2011-05-14 16:00:52 +00:00
|
|
|
return ctxt->ops->read_std(ctxt, addr, desc, sizeof *desc,
|
|
|
|
|
&ctxt->exception);
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
2010-07-29 12:11:51 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
/* allowed just for 8 bytes segments */
|
|
|
|
|
static int write_segment_descriptor(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 selector, struct desc_struct *desc)
|
|
|
|
|
{
|
|
|
|
|
struct desc_ptr dt;
|
|
|
|
|
u16 index = selector >> 3;
|
|
|
|
|
ulong addr;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
get_descriptor_table_ptr(ctxt, selector, &dt);
|
2007-12-06 16:14:14 +00:00
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
if (dt.size < index * 8 + 7)
|
|
|
|
|
return emulate_gp(ctxt, selector & 0xfffc);
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
addr = dt.address + index * 8;
|
2011-05-14 16:00:52 +00:00
|
|
|
return ctxt->ops->write_std(ctxt, addr, desc, sizeof *desc,
|
|
|
|
|
&ctxt->exception);
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
2007-10-28 14:34:25 +00:00
|
|
|
|
2011-03-07 12:55:06 +00:00
|
|
|
/* Does not support long mode */
|
2010-07-29 12:11:52 +00:00
|
|
|
static int load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 selector, int seg)
|
|
|
|
|
{
|
2012-06-13 13:30:53 +00:00
|
|
|
struct desc_struct seg_desc, old_desc;
|
2010-07-29 12:11:52 +00:00
|
|
|
u8 dpl, rpl, cpl;
|
|
|
|
|
unsigned err_vec = GP_VECTOR;
|
|
|
|
|
u32 err_code = 0;
|
|
|
|
|
bool null_selector = !(selector & ~0x3); /* 0000-0003 are null */
|
2012-06-13 13:29:39 +00:00
|
|
|
ulong desc_addr;
|
2010-07-29 12:11:52 +00:00
|
|
|
int ret;
|
2012-08-21 14:07:04 +00:00
|
|
|
u16 dummy;
|
2010-03-18 13:20:20 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
memset(&seg_desc, 0, sizeof seg_desc);
|
2010-03-18 13:20:20 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if ((seg <= VCPU_SREG_GS && ctxt->mode == X86EMUL_MODE_VM86)
|
|
|
|
|
|| ctxt->mode == X86EMUL_MODE_REAL) {
|
|
|
|
|
/* set real mode segment descriptor */
|
2012-08-21 14:07:04 +00:00
|
|
|
ctxt->ops->get_segment(ctxt, &dummy, &seg_desc, NULL, seg);
|
2010-07-29 12:11:52 +00:00
|
|
|
set_desc_base(&seg_desc, selector << 4);
|
|
|
|
|
goto load;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-07 14:03:42 +00:00
|
|
|
rpl = selector & 3;
|
|
|
|
|
cpl = ctxt->ops->cpl(ctxt);
|
|
|
|
|
|
|
|
|
|
/* NULL selector is not valid for TR, CS and SS (except for long mode) */
|
|
|
|
|
if ((seg == VCPU_SREG_CS
|
|
|
|
|
|| (seg == VCPU_SREG_SS
|
|
|
|
|
&& (ctxt->mode != X86EMUL_MODE_PROT64 || rpl != cpl))
|
|
|
|
|
|| seg == VCPU_SREG_TR)
|
2010-07-29 12:11:52 +00:00
|
|
|
&& null_selector)
|
|
|
|
|
goto exception;
|
|
|
|
|
|
|
|
|
|
/* TR should be in GDT only */
|
|
|
|
|
if (seg == VCPU_SREG_TR && (selector & (1 << 2)))
|
|
|
|
|
goto exception;
|
|
|
|
|
|
|
|
|
|
if (null_selector) /* for NULL selector skip all following checks */
|
|
|
|
|
goto load;
|
|
|
|
|
|
2012-06-13 13:29:39 +00:00
|
|
|
ret = read_segment_descriptor(ctxt, selector, &seg_desc, &desc_addr);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
err_code = selector & 0xfffc;
|
|
|
|
|
err_vec = GP_VECTOR;
|
|
|
|
|
|
2012-06-28 07:19:51 +00:00
|
|
|
/* can't load system descriptor into segment selector */
|
2010-07-29 12:11:52 +00:00
|
|
|
if (seg <= VCPU_SREG_GS && !seg_desc.s)
|
|
|
|
|
goto exception;
|
|
|
|
|
|
|
|
|
|
if (!seg_desc.p) {
|
|
|
|
|
err_vec = (seg == VCPU_SREG_SS) ? SS_VECTOR : NP_VECTOR;
|
|
|
|
|
goto exception;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
dpl = seg_desc.dpl;
|
|
|
|
|
|
|
|
|
|
switch (seg) {
|
|
|
|
|
case VCPU_SREG_SS:
|
|
|
|
|
/*
|
|
|
|
|
* segment is not a writable data segment or segment
|
|
|
|
|
* selector's RPL != CPL or segment selector's RPL != CPL
|
|
|
|
|
*/
|
|
|
|
|
if (rpl != cpl || (seg_desc.type & 0xa) != 0x2 || dpl != cpl)
|
|
|
|
|
goto exception;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2010-07-29 12:11:52 +00:00
|
|
|
case VCPU_SREG_CS:
|
|
|
|
|
if (!(seg_desc.type & 8))
|
|
|
|
|
goto exception;
|
|
|
|
|
|
|
|
|
|
if (seg_desc.type & 4) {
|
|
|
|
|
/* conforming */
|
|
|
|
|
if (dpl > cpl)
|
|
|
|
|
goto exception;
|
|
|
|
|
} else {
|
|
|
|
|
/* nonconforming */
|
|
|
|
|
if (rpl > cpl || dpl != cpl)
|
|
|
|
|
goto exception;
|
|
|
|
|
}
|
|
|
|
|
/* CS(RPL) <- CPL */
|
|
|
|
|
selector = (selector & 0xfffc) | cpl;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2010-07-29 12:11:52 +00:00
|
|
|
case VCPU_SREG_TR:
|
|
|
|
|
if (seg_desc.s || (seg_desc.type != 1 && seg_desc.type != 9))
|
|
|
|
|
goto exception;
|
2012-06-13 13:30:53 +00:00
|
|
|
old_desc = seg_desc;
|
|
|
|
|
seg_desc.type |= 2; /* busy */
|
|
|
|
|
ret = ctxt->ops->cmpxchg_emulated(ctxt, desc_addr, &old_desc, &seg_desc,
|
|
|
|
|
sizeof(seg_desc), &ctxt->exception);
|
|
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
case VCPU_SREG_LDTR:
|
|
|
|
|
if (seg_desc.s || seg_desc.type != 2)
|
|
|
|
|
goto exception;
|
|
|
|
|
break;
|
|
|
|
|
default: /* DS, ES, FS, or GS */
|
2007-10-17 17:30:41 +00:00
|
|
|
/*
|
2010-07-29 12:11:52 +00:00
|
|
|
* segment is not a data or readable code segment or
|
|
|
|
|
* ((segment is a data or nonconforming code segment)
|
|
|
|
|
* and (both RPL and CPL > DPL))
|
2007-10-17 17:30:41 +00:00
|
|
|
*/
|
2010-07-29 12:11:52 +00:00
|
|
|
if ((seg_desc.type & 0xa) == 0x8 ||
|
|
|
|
|
(((seg_desc.type & 0xc) != 0xc) &&
|
|
|
|
|
(rpl > dpl && cpl > dpl)))
|
|
|
|
|
goto exception;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (seg_desc.s) {
|
|
|
|
|
/* mark segment as accessed */
|
|
|
|
|
seg_desc.type |= 1;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = write_segment_descriptor(ctxt, selector, &seg_desc);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
load:
|
2011-05-14 16:00:52 +00:00
|
|
|
ctxt->ops->set_segment(ctxt, selector, &seg_desc, 0, seg);
|
2010-07-29 12:11:52 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
exception:
|
|
|
|
|
emulate_exception(ctxt, err_vec, err_code, true);
|
|
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-17 01:17:30 +00:00
|
|
|
static void write_register_operand(struct operand *op)
|
|
|
|
|
{
|
|
|
|
|
/* The 4-byte case *is* correct: in 64-bit mode we zero-extend. */
|
|
|
|
|
switch (op->bytes) {
|
|
|
|
|
case 1:
|
|
|
|
|
*(u8 *)op->addr.reg = (u8)op->val;
|
|
|
|
|
break;
|
|
|
|
|
case 2:
|
|
|
|
|
*(u16 *)op->addr.reg = (u16)op->val;
|
|
|
|
|
break;
|
|
|
|
|
case 4:
|
|
|
|
|
*op->addr.reg = (u32)op->val;
|
|
|
|
|
break; /* 64b: zero-extend */
|
|
|
|
|
case 8:
|
|
|
|
|
*op->addr.reg = op->val;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-01 17:26:23 +00:00
|
|
|
static int writeback(struct x86_emulate_ctxt *ctxt)
|
2010-07-29 12:11:52 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->dst.type) {
|
2010-07-29 12:11:52 +00:00
|
|
|
case OP_REG:
|
2011-06-01 12:34:25 +00:00
|
|
|
write_register_operand(&ctxt->dst);
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2010-07-29 12:11:52 +00:00
|
|
|
case OP_MEM:
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->lock_prefix)
|
2011-03-31 14:52:26 +00:00
|
|
|
rc = segmented_cmpxchg(ctxt,
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.addr.mem,
|
|
|
|
|
&ctxt->dst.orig_val,
|
|
|
|
|
&ctxt->dst.val,
|
|
|
|
|
ctxt->dst.bytes);
|
2009-04-12 10:36:41 +00:00
|
|
|
else
|
2011-03-31 14:52:26 +00:00
|
|
|
rc = segmented_write(ctxt,
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.addr.mem,
|
|
|
|
|
&ctxt->dst.val,
|
|
|
|
|
ctxt->dst.bytes);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-03-18 13:20:21 +00:00
|
|
|
break;
|
2011-03-29 09:41:27 +00:00
|
|
|
case OP_XMM:
|
2011-06-01 12:34:25 +00:00
|
|
|
write_sse_reg(ctxt, &ctxt->dst.vec_val, ctxt->dst.addr.xmm);
|
2011-03-29 09:41:27 +00:00
|
|
|
break;
|
2012-04-09 15:40:02 +00:00
|
|
|
case OP_MM:
|
|
|
|
|
write_mmx_reg(ctxt, &ctxt->dst.mm_val, ctxt->dst.addr.mm);
|
|
|
|
|
break;
|
2010-07-29 12:11:52 +00:00
|
|
|
case OP_NONE:
|
|
|
|
|
/* no writeback */
|
2010-04-28 16:15:26 +00:00
|
|
|
break;
|
2010-07-29 12:11:52 +00:00
|
|
|
default:
|
2010-04-28 16:15:26 +00:00
|
|
|
break;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
}
|
2010-07-29 12:11:52 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
2012-06-12 17:19:40 +00:00
|
|
|
static int push(struct x86_emulate_ctxt *ctxt, void *data, int bytes)
|
2010-07-29 12:11:52 +00:00
|
|
|
{
|
2011-04-12 15:29:09 +00:00
|
|
|
struct segmented_address addr;
|
2008-12-04 13:26:42 +00:00
|
|
|
|
2012-08-19 11:34:31 +00:00
|
|
|
rsp_increment(ctxt, -bytes);
|
2012-08-27 20:46:17 +00:00
|
|
|
addr.ea = reg_read(ctxt, VCPU_REGS_RSP) & stack_mask(ctxt);
|
2011-04-12 15:29:09 +00:00
|
|
|
addr.seg = VCPU_SREG_SS;
|
|
|
|
|
|
2012-06-12 17:19:40 +00:00
|
|
|
return segmented_write(ctxt, addr, data, bytes);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_push(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-04-12 15:29:09 +00:00
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
2012-06-12 17:19:40 +00:00
|
|
|
return push(ctxt, &ctxt->src.val, ctxt->op_bytes);
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
2010-03-18 13:20:20 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
static int emulate_pop(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
void *dest, int len)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
2010-11-17 13:28:21 +00:00
|
|
|
struct segmented_address addr;
|
2007-09-18 09:27:19 +00:00
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
addr.ea = reg_read(ctxt, VCPU_REGS_RSP) & stack_mask(ctxt);
|
2010-11-17 13:28:21 +00:00
|
|
|
addr.seg = VCPU_SREG_SS;
|
2011-03-31 14:52:26 +00:00
|
|
|
rc = segmented_read(ctxt, addr, dest, len);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2012-08-19 11:34:31 +00:00
|
|
|
rsp_increment(ctxt, len);
|
2010-07-29 12:11:52 +00:00
|
|
|
return rc;
|
2007-09-18 09:27:19 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-23 09:49:40 +00:00
|
|
|
static int em_pop(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
return emulate_pop(ctxt, &ctxt->dst.val, ctxt->op_bytes);
|
2011-04-23 09:49:40 +00:00
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
static int emulate_popf(struct x86_emulate_ctxt *ctxt,
|
2011-05-14 16:00:52 +00:00
|
|
|
void *dest, int len)
|
2010-04-28 16:15:22 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
2010-07-29 12:11:52 +00:00
|
|
|
unsigned long val, change_mask;
|
|
|
|
|
int iopl = (ctxt->eflags & X86_EFLAGS_IOPL) >> IOPL_SHIFT;
|
2011-05-14 16:00:52 +00:00
|
|
|
int cpl = ctxt->ops->cpl(ctxt);
|
2010-04-28 16:15:22 +00:00
|
|
|
|
2011-05-01 17:27:55 +00:00
|
|
|
rc = emulate_pop(ctxt, &val, len);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-04-28 16:15:22 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
change_mask = EFLG_CF | EFLG_PF | EFLG_AF | EFLG_ZF | EFLG_SF | EFLG_OF
|
|
|
|
|
| EFLG_TF | EFLG_DF | EFLG_NT | EFLG_RF | EFLG_AC | EFLG_ID;
|
2010-04-28 16:15:22 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
switch(ctxt->mode) {
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
|
|
|
|
case X86EMUL_MODE_PROT32:
|
|
|
|
|
case X86EMUL_MODE_PROT16:
|
|
|
|
|
if (cpl == 0)
|
|
|
|
|
change_mask |= EFLG_IOPL;
|
|
|
|
|
if (cpl <= iopl)
|
|
|
|
|
change_mask |= EFLG_IF;
|
|
|
|
|
break;
|
|
|
|
|
case X86EMUL_MODE_VM86:
|
2010-11-22 15:53:25 +00:00
|
|
|
if (iopl < 3)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2010-07-29 12:11:52 +00:00
|
|
|
change_mask |= EFLG_IF;
|
|
|
|
|
break;
|
|
|
|
|
default: /* real mode */
|
|
|
|
|
change_mask |= (EFLG_IOPL | EFLG_IF);
|
|
|
|
|
break;
|
2010-04-28 16:15:22 +00:00
|
|
|
}
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
*(unsigned long *)dest =
|
|
|
|
|
(ctxt->eflags & ~change_mask) | (val & change_mask);
|
|
|
|
|
|
|
|
|
|
return rc;
|
2010-04-28 16:15:22 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-23 09:52:56 +00:00
|
|
|
static int em_popf(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_REG;
|
|
|
|
|
ctxt->dst.addr.reg = &ctxt->eflags;
|
|
|
|
|
ctxt->dst.bytes = ctxt->op_bytes;
|
|
|
|
|
return emulate_popf(ctxt, &ctxt->dst.val, ctxt->op_bytes);
|
2011-04-23 09:52:56 +00:00
|
|
|
}
|
|
|
|
|
|
2012-06-12 17:03:23 +00:00
|
|
|
static int em_enter(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
unsigned frame_size = ctxt->src.val;
|
|
|
|
|
unsigned nesting_level = ctxt->src2.val & 31;
|
2012-08-27 20:46:17 +00:00
|
|
|
ulong rbp;
|
2012-06-12 17:03:23 +00:00
|
|
|
|
|
|
|
|
if (nesting_level)
|
|
|
|
|
return X86EMUL_UNHANDLEABLE;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
rbp = reg_read(ctxt, VCPU_REGS_RBP);
|
|
|
|
|
rc = push(ctxt, &rbp, stack_size(ctxt));
|
2012-06-12 17:03:23 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2012-08-27 20:46:17 +00:00
|
|
|
assign_masked(reg_rmw(ctxt, VCPU_REGS_RBP), reg_read(ctxt, VCPU_REGS_RSP),
|
2012-06-12 17:03:23 +00:00
|
|
|
stack_mask(ctxt));
|
2012-08-27 20:46:17 +00:00
|
|
|
assign_masked(reg_rmw(ctxt, VCPU_REGS_RSP),
|
|
|
|
|
reg_read(ctxt, VCPU_REGS_RSP) - frame_size,
|
2012-06-12 17:03:23 +00:00
|
|
|
stack_mask(ctxt));
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-07 14:49:24 +00:00
|
|
|
static int em_leave(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
assign_masked(reg_rmw(ctxt, VCPU_REGS_RSP), reg_read(ctxt, VCPU_REGS_RBP),
|
2012-06-07 14:49:24 +00:00
|
|
|
stack_mask(ctxt));
|
2012-08-27 20:46:17 +00:00
|
|
|
return emulate_pop(ctxt, reg_rmw(ctxt, VCPU_REGS_RBP), ctxt->op_bytes);
|
2012-06-07 14:49:24 +00:00
|
|
|
}
|
|
|
|
|
|
2011-09-13 07:45:51 +00:00
|
|
|
static int em_push_sreg(struct x86_emulate_ctxt *ctxt)
|
2010-03-18 13:20:27 +00:00
|
|
|
{
|
2011-09-13 07:45:51 +00:00
|
|
|
int seg = ctxt->src2.val;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = get_segment_selector(ctxt, seg);
|
2010-03-18 13:20:27 +00:00
|
|
|
|
2011-04-12 15:31:23 +00:00
|
|
|
return em_push(ctxt);
|
2010-03-18 13:20:27 +00:00
|
|
|
}
|
|
|
|
|
|
2011-09-13 07:45:51 +00:00
|
|
|
static int em_pop_sreg(struct x86_emulate_ctxt *ctxt)
|
2010-03-18 13:20:17 +00:00
|
|
|
{
|
2011-09-13 07:45:51 +00:00
|
|
|
int seg = ctxt->src2.val;
|
2010-07-29 12:11:52 +00:00
|
|
|
unsigned long selector;
|
|
|
|
|
int rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = emulate_pop(ctxt, &selector, ctxt->op_bytes);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
rc = load_segment_descriptor(ctxt, (u16)selector, seg);
|
2010-07-29 12:11:52 +00:00
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-23 09:51:07 +00:00
|
|
|
static int em_pusha(struct x86_emulate_ctxt *ctxt)
|
2010-03-18 13:20:17 +00:00
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
unsigned long old_esp = reg_read(ctxt, VCPU_REGS_RSP);
|
2010-07-29 12:11:52 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
|
|
|
|
int reg = VCPU_REGS_RAX;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
while (reg <= VCPU_REGS_RDI) {
|
|
|
|
|
(reg == VCPU_REGS_RSP) ?
|
2012-08-27 20:46:17 +00:00
|
|
|
(ctxt->src.val = old_esp) : (ctxt->src.val = reg_read(ctxt, reg));
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
++reg;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-23 09:52:56 +00:00
|
|
|
static int em_pushf(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = (unsigned long)ctxt->eflags;
|
2011-04-23 09:52:56 +00:00
|
|
|
return em_push(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-23 09:51:07 +00:00
|
|
|
static int em_popa(struct x86_emulate_ctxt *ctxt)
|
2010-03-18 13:20:17 +00:00
|
|
|
{
|
2010-07-29 12:11:52 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
|
|
|
|
int reg = VCPU_REGS_RDI;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
while (reg >= VCPU_REGS_RAX) {
|
|
|
|
|
if (reg == VCPU_REGS_RSP) {
|
2012-08-19 11:34:31 +00:00
|
|
|
rsp_increment(ctxt, ctxt->op_bytes);
|
2010-07-29 12:11:52 +00:00
|
|
|
--reg;
|
|
|
|
|
}
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
rc = emulate_pop(ctxt, reg_rmw(ctxt, reg), ctxt->op_bytes);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
break;
|
|
|
|
|
--reg;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
2010-07-29 12:11:52 +00:00
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
static int __emulate_int_real(struct x86_emulate_ctxt *ctxt, int irq)
|
2010-08-04 11:38:06 +00:00
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-08-17 08:17:51 +00:00
|
|
|
int rc;
|
2010-08-04 11:38:06 +00:00
|
|
|
struct desc_ptr dt;
|
|
|
|
|
gva_t cs_addr;
|
|
|
|
|
gva_t eip_addr;
|
|
|
|
|
u16 cs, eip;
|
|
|
|
|
|
|
|
|
|
/* TODO: Add limit checks */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = ctxt->eflags;
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2010-08-17 08:17:51 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-08-04 11:38:06 +00:00
|
|
|
|
|
|
|
|
ctxt->eflags &= ~(EFLG_IF | EFLG_TF | EFLG_AC);
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = get_segment_selector(ctxt, VCPU_SREG_CS);
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2010-08-17 08:17:51 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-08-04 11:38:06 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = ctxt->_eip;
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2010-08-17 08:17:51 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_idt(ctxt, &dt);
|
2010-08-04 11:38:06 +00:00
|
|
|
|
|
|
|
|
eip_addr = dt.address + (irq << 2);
|
|
|
|
|
cs_addr = dt.address + (irq << 2) + 2;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
rc = ops->read_std(ctxt, cs_addr, &cs, 2, &ctxt->exception);
|
2010-08-04 11:38:06 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
rc = ops->read_std(ctxt, eip_addr, &eip, 2, &ctxt->exception);
|
2010-08-04 11:38:06 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
rc = load_segment_descriptor(ctxt, cs, VCPU_SREG_CS);
|
2010-08-04 11:38:06 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = eip;
|
2010-08-04 11:38:06 +00:00
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
int emulate_int_real(struct x86_emulate_ctxt *ctxt, int irq)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
invalidate_registers(ctxt);
|
|
|
|
|
rc = __emulate_int_real(ctxt, irq);
|
|
|
|
|
if (rc == X86EMUL_CONTINUE)
|
|
|
|
|
writeback_registers(ctxt);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
static int emulate_int(struct x86_emulate_ctxt *ctxt, int irq)
|
2010-08-04 11:38:06 +00:00
|
|
|
{
|
|
|
|
|
switch(ctxt->mode) {
|
|
|
|
|
case X86EMUL_MODE_REAL:
|
2012-08-27 20:46:17 +00:00
|
|
|
return __emulate_int_real(ctxt, irq);
|
2010-08-04 11:38:06 +00:00
|
|
|
case X86EMUL_MODE_VM86:
|
|
|
|
|
case X86EMUL_MODE_PROT16:
|
|
|
|
|
case X86EMUL_MODE_PROT32:
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
|
|
|
|
default:
|
|
|
|
|
/* Protected mode interrupts unimplemented yet */
|
|
|
|
|
return X86EMUL_UNHANDLEABLE;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
static int emulate_iret_real(struct x86_emulate_ctxt *ctxt)
|
2010-03-18 13:20:17 +00:00
|
|
|
{
|
2010-07-29 12:11:52 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
|
|
|
|
unsigned long temp_eip = 0;
|
|
|
|
|
unsigned long temp_eflags = 0;
|
|
|
|
|
unsigned long cs = 0;
|
|
|
|
|
unsigned long mask = EFLG_CF | EFLG_PF | EFLG_AF | EFLG_ZF | EFLG_SF | EFLG_TF |
|
|
|
|
|
EFLG_IF | EFLG_DF | EFLG_OF | EFLG_IOPL | EFLG_NT | EFLG_RF |
|
|
|
|
|
EFLG_AC | EFLG_ID | (1 << 1); /* Last one is the reserved bit */
|
|
|
|
|
unsigned long vm86_mask = EFLG_VM | EFLG_VIF | EFLG_VIP;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
/* TODO: Add stack limit check */
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = emulate_pop(ctxt, &temp_eip, ctxt->op_bytes);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-11-22 15:53:25 +00:00
|
|
|
if (temp_eip & ~0xffff)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = emulate_pop(ctxt, &cs, ctxt->op_bytes);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = emulate_pop(ctxt, &temp_eflags, ctxt->op_bytes);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
rc = load_segment_descriptor(ctxt, (u16)cs, VCPU_SREG_CS);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = temp_eip;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->op_bytes == 4)
|
2010-07-29 12:11:52 +00:00
|
|
|
ctxt->eflags = ((temp_eflags & mask) | (ctxt->eflags & vm86_mask));
|
2011-06-01 12:34:25 +00:00
|
|
|
else if (ctxt->op_bytes == 2) {
|
2010-07-29 12:11:52 +00:00
|
|
|
ctxt->eflags &= ~0xffff;
|
|
|
|
|
ctxt->eflags |= temp_eflags;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
ctxt->eflags &= ~EFLG_RESERVED_ZEROS_MASK; /* Clear reserved zeros */
|
|
|
|
|
ctxt->eflags |= EFLG_RESERVED_ONE_MASK;
|
|
|
|
|
|
|
|
|
|
return rc;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:55:10 +00:00
|
|
|
static int em_iret(struct x86_emulate_ctxt *ctxt)
|
2010-06-15 01:03:33 +00:00
|
|
|
{
|
2010-07-29 12:11:52 +00:00
|
|
|
switch(ctxt->mode) {
|
|
|
|
|
case X86EMUL_MODE_REAL:
|
2011-05-14 16:00:52 +00:00
|
|
|
return emulate_iret_real(ctxt);
|
2010-07-29 12:11:52 +00:00
|
|
|
case X86EMUL_MODE_VM86:
|
|
|
|
|
case X86EMUL_MODE_PROT16:
|
|
|
|
|
case X86EMUL_MODE_PROT32:
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
2010-06-15 01:03:33 +00:00
|
|
|
default:
|
2010-07-29 12:11:52 +00:00
|
|
|
/* iret from protected mode unimplemented yet */
|
|
|
|
|
return X86EMUL_UNHANDLEABLE;
|
2010-06-15 01:03:33 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-01 17:30:48 +00:00
|
|
|
static int em_jmp_far(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
unsigned short sel;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
memcpy(&sel, ctxt->src.valptr + ctxt->op_bytes, 2);
|
2011-05-01 17:30:48 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
rc = load_segment_descriptor(ctxt, sel, VCPU_SREG_CS);
|
2011-05-01 17:30:48 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = 0;
|
|
|
|
|
memcpy(&ctxt->_eip, ctxt->src.valptr, ctxt->op_bytes);
|
2011-05-01 17:30:48 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-01 17:29:17 +00:00
|
|
|
static int em_grp2(struct x86_emulate_ctxt *ctxt)
|
2007-09-24 09:10:54 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->modrm_reg) {
|
2007-09-24 09:10:54 +00:00
|
|
|
case 0: /* rol */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "rol");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 1: /* ror */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "ror");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 2: /* rcl */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "rcl");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 3: /* rcr */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "rcr");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 4: /* sal/shl */
|
|
|
|
|
case 6: /* sal/shl */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "sal");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 5: /* shr */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "shr");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 7: /* sar */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcB(ctxt, "sar");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2011-05-01 17:29:17 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2007-09-24 09:10:54 +00:00
|
|
|
}
|
|
|
|
|
|
2011-09-13 07:45:39 +00:00
|
|
|
static int em_not(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
ctxt->dst.val = ~ctxt->dst.val;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_neg(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
emulate_1op(ctxt, "neg");
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_mul_ex(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u8 ex = 0;
|
|
|
|
|
|
|
|
|
|
emulate_1op_rax_rdx(ctxt, "mul", ex);
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_imul_ex(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u8 ex = 0;
|
|
|
|
|
|
|
|
|
|
emulate_1op_rax_rdx(ctxt, "imul", ex);
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_div_ex(struct x86_emulate_ctxt *ctxt)
|
2007-09-24 09:10:54 +00:00
|
|
|
{
|
2010-08-26 08:59:01 +00:00
|
|
|
u8 de = 0;
|
2007-09-24 09:10:54 +00:00
|
|
|
|
2011-09-13 07:45:39 +00:00
|
|
|
emulate_1op_rax_rdx(ctxt, "div", de);
|
|
|
|
|
if (de)
|
|
|
|
|
return emulate_de(ctxt);
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_idiv_ex(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u8 de = 0;
|
|
|
|
|
|
|
|
|
|
emulate_1op_rax_rdx(ctxt, "idiv", de);
|
2010-08-26 08:59:01 +00:00
|
|
|
if (de)
|
|
|
|
|
return emulate_de(ctxt);
|
2010-08-08 18:11:38 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2007-09-24 09:10:54 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-01 17:29:17 +00:00
|
|
|
static int em_grp45(struct x86_emulate_ctxt *ctxt)
|
2007-09-24 09:10:54 +00:00
|
|
|
{
|
2011-04-12 15:29:09 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
2007-09-24 09:10:54 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->modrm_reg) {
|
2007-09-24 09:10:54 +00:00
|
|
|
case 0: /* inc */
|
2011-09-07 13:41:38 +00:00
|
|
|
emulate_1op(ctxt, "inc");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
case 1: /* dec */
|
2011-09-07 13:41:38 +00:00
|
|
|
emulate_1op(ctxt, "dec");
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
2008-09-08 18:47:19 +00:00
|
|
|
case 2: /* call near abs */ {
|
|
|
|
|
long int old_eip;
|
2011-06-01 12:34:25 +00:00
|
|
|
old_eip = ctxt->_eip;
|
|
|
|
|
ctxt->_eip = ctxt->src.val;
|
|
|
|
|
ctxt->src.val = old_eip;
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2008-09-08 18:47:19 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2007-09-24 09:10:54 +00:00
|
|
|
case 4: /* jmp abs */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = ctxt->src.val;
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
2011-05-01 17:30:48 +00:00
|
|
|
case 5: /* jmp far */
|
|
|
|
|
rc = em_jmp_far(ctxt);
|
|
|
|
|
break;
|
2007-09-24 09:10:54 +00:00
|
|
|
case 6: /* push */
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2007-09-24 09:10:54 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2011-04-12 15:29:09 +00:00
|
|
|
return rc;
|
2007-09-24 09:10:54 +00:00
|
|
|
}
|
|
|
|
|
|
2011-12-06 09:07:27 +00:00
|
|
|
static int em_cmpxchg8b(struct x86_emulate_ctxt *ctxt)
|
2007-09-24 09:10:54 +00:00
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
u64 old = ctxt->dst.orig_val64;
|
2007-09-24 09:10:54 +00:00
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
if (((u32) (old >> 0) != (u32) reg_read(ctxt, VCPU_REGS_RAX)) ||
|
|
|
|
|
((u32) (old >> 32) != (u32) reg_read(ctxt, VCPU_REGS_RDX))) {
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = (u32) (old >> 0);
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = (u32) (old >> 32);
|
2007-09-24 09:10:55 +00:00
|
|
|
ctxt->eflags &= ~EFLG_ZF;
|
2007-09-24 09:10:54 +00:00
|
|
|
} else {
|
2012-08-27 20:46:17 +00:00
|
|
|
ctxt->dst.val64 = ((u64)reg_read(ctxt, VCPU_REGS_RCX) << 32) |
|
|
|
|
|
(u32) reg_read(ctxt, VCPU_REGS_RBX);
|
2007-09-24 09:10:54 +00:00
|
|
|
|
2007-09-24 09:10:55 +00:00
|
|
|
ctxt->eflags |= EFLG_ZF;
|
2007-09-24 09:10:54 +00:00
|
|
|
}
|
2010-02-12 06:57:56 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2007-09-24 09:10:54 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-29 13:00:22 +00:00
|
|
|
static int em_ret(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_REG;
|
|
|
|
|
ctxt->dst.addr.reg = &ctxt->_eip;
|
|
|
|
|
ctxt->dst.bytes = ctxt->op_bytes;
|
2011-05-29 13:00:22 +00:00
|
|
|
return em_pop(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:55:10 +00:00
|
|
|
static int em_ret_far(struct x86_emulate_ctxt *ctxt)
|
2009-01-05 11:27:34 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
unsigned long cs;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = emulate_pop(ctxt, &ctxt->_eip, ctxt->op_bytes);
|
2010-02-12 06:57:56 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
2009-01-05 11:27:34 +00:00
|
|
|
return rc;
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->op_bytes == 4)
|
|
|
|
|
ctxt->_eip = (u32)ctxt->_eip;
|
|
|
|
|
rc = emulate_pop(ctxt, &cs, ctxt->op_bytes);
|
2010-02-12 06:57:56 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
2009-01-05 11:27:34 +00:00
|
|
|
return rc;
|
2011-05-14 16:00:52 +00:00
|
|
|
rc = load_segment_descriptor(ctxt, (u16)cs, VCPU_SREG_CS);
|
2009-01-05 11:27:34 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:20:47 +00:00
|
|
|
static int em_cmpxchg(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
/* Save real source value, then compare EAX against destination. */
|
|
|
|
|
ctxt->src.orig_val = ctxt->src.val;
|
2012-08-27 20:46:17 +00:00
|
|
|
ctxt->src.val = reg_read(ctxt, VCPU_REGS_RAX);
|
2011-11-22 06:20:47 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "cmp");
|
|
|
|
|
|
|
|
|
|
if (ctxt->eflags & EFLG_ZF) {
|
|
|
|
|
/* Success: write back to memory. */
|
|
|
|
|
ctxt->dst.val = ctxt->src.orig_val;
|
|
|
|
|
} else {
|
|
|
|
|
/* Failure: write the value we saw to EAX. */
|
|
|
|
|
ctxt->dst.type = OP_REG;
|
2012-08-27 20:46:17 +00:00
|
|
|
ctxt->dst.addr.reg = reg_rmw(ctxt, VCPU_REGS_RAX);
|
2011-11-22 06:20:47 +00:00
|
|
|
}
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-09-13 07:45:50 +00:00
|
|
|
static int em_lseg(struct x86_emulate_ctxt *ctxt)
|
2010-08-23 06:56:54 +00:00
|
|
|
{
|
2011-09-13 07:45:50 +00:00
|
|
|
int seg = ctxt->src2.val;
|
2010-08-23 06:56:54 +00:00
|
|
|
unsigned short sel;
|
|
|
|
|
int rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
memcpy(&sel, ctxt->src.valptr + ctxt->op_bytes, 2);
|
2010-08-23 06:56:54 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
rc = load_segment_descriptor(ctxt, sel, seg);
|
2010-08-23 06:56:54 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ctxt->src.val;
|
2010-08-23 06:56:54 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
static void
|
2009-06-18 10:56:00 +00:00
|
|
|
setup_syscalls_segments(struct x86_emulate_ctxt *ctxt,
|
2011-05-14 16:00:52 +00:00
|
|
|
struct desc_struct *cs, struct desc_struct *ss)
|
2009-06-18 10:56:00 +00:00
|
|
|
{
|
|
|
|
|
cs->l = 0; /* will be adjusted later */
|
2010-04-28 16:15:30 +00:00
|
|
|
set_desc_base(cs, 0); /* flat segment */
|
2009-06-18 10:56:00 +00:00
|
|
|
cs->g = 1; /* 4kb granularity */
|
2010-04-28 16:15:30 +00:00
|
|
|
set_desc_limit(cs, 0xfffff); /* 4GB limit */
|
2009-06-18 10:56:00 +00:00
|
|
|
cs->type = 0x0b; /* Read, Execute, Accessed */
|
|
|
|
|
cs->s = 1;
|
|
|
|
|
cs->dpl = 0; /* will be adjusted later */
|
2010-04-28 16:15:30 +00:00
|
|
|
cs->p = 1;
|
|
|
|
|
cs->d = 1;
|
2012-07-25 12:49:42 +00:00
|
|
|
cs->avl = 0;
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2010-04-28 16:15:30 +00:00
|
|
|
set_desc_base(ss, 0); /* flat segment */
|
|
|
|
|
set_desc_limit(ss, 0xfffff); /* 4GB limit */
|
2009-06-18 10:56:00 +00:00
|
|
|
ss->g = 1; /* 4kb granularity */
|
|
|
|
|
ss->s = 1;
|
|
|
|
|
ss->type = 0x03; /* Read/Write, Accessed */
|
2010-04-28 16:15:30 +00:00
|
|
|
ss->d = 1; /* 32bit stack segment */
|
2009-06-18 10:56:00 +00:00
|
|
|
ss->dpl = 0;
|
2010-04-28 16:15:30 +00:00
|
|
|
ss->p = 1;
|
2012-07-25 12:49:42 +00:00
|
|
|
ss->l = 0;
|
|
|
|
|
ss->avl = 0;
|
2009-06-18 10:56:00 +00:00
|
|
|
}
|
|
|
|
|
|
2012-02-01 10:23:21 +00:00
|
|
|
static bool vendor_intel(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u32 eax, ebx, ecx, edx;
|
|
|
|
|
|
|
|
|
|
eax = ecx = 0;
|
2012-06-07 11:10:16 +00:00
|
|
|
ctxt->ops->get_cpuid(ctxt, &eax, &ebx, &ecx, &edx);
|
|
|
|
|
return ebx == X86EMUL_CPUID_VENDOR_GenuineIntel_ebx
|
2012-02-01 10:23:21 +00:00
|
|
|
&& ecx == X86EMUL_CPUID_VENDOR_GenuineIntel_ecx
|
|
|
|
|
&& edx == X86EMUL_CPUID_VENDOR_GenuineIntel_edx;
|
|
|
|
|
}
|
|
|
|
|
|
2012-01-12 15:43:04 +00:00
|
|
|
static bool em_syscall_is_enabled(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
|
|
|
|
u32 eax, ebx, ecx, edx;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* syscall should always be enabled in longmode - so only become
|
|
|
|
|
* vendor specific (cpuid) if other modes are active...
|
|
|
|
|
*/
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
eax = 0x00000000;
|
|
|
|
|
ecx = 0x00000000;
|
2012-06-07 11:10:16 +00:00
|
|
|
ops->get_cpuid(ctxt, &eax, &ebx, &ecx, &edx);
|
|
|
|
|
/*
|
|
|
|
|
* Intel ("GenuineIntel")
|
|
|
|
|
* remark: Intel CPUs only support "syscall" in 64bit
|
|
|
|
|
* longmode. Also an 64bit guest with a
|
|
|
|
|
* 32bit compat-app running will #UD !! While this
|
|
|
|
|
* behaviour can be fixed (by emulating) into AMD
|
|
|
|
|
* response - CPUs of AMD can't behave like Intel.
|
|
|
|
|
*/
|
|
|
|
|
if (ebx == X86EMUL_CPUID_VENDOR_GenuineIntel_ebx &&
|
|
|
|
|
ecx == X86EMUL_CPUID_VENDOR_GenuineIntel_ecx &&
|
|
|
|
|
edx == X86EMUL_CPUID_VENDOR_GenuineIntel_edx)
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
/* AMD ("AuthenticAMD") */
|
|
|
|
|
if (ebx == X86EMUL_CPUID_VENDOR_AuthenticAMD_ebx &&
|
|
|
|
|
ecx == X86EMUL_CPUID_VENDOR_AuthenticAMD_ecx &&
|
|
|
|
|
edx == X86EMUL_CPUID_VENDOR_AuthenticAMD_edx)
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
/* AMD ("AMDisbetter!") */
|
|
|
|
|
if (ebx == X86EMUL_CPUID_VENDOR_AMDisbetterI_ebx &&
|
|
|
|
|
ecx == X86EMUL_CPUID_VENDOR_AMDisbetterI_ecx &&
|
|
|
|
|
edx == X86EMUL_CPUID_VENDOR_AMDisbetterI_edx)
|
|
|
|
|
return true;
|
2012-01-12 15:43:04 +00:00
|
|
|
|
|
|
|
|
/* default: (not Intel, not AMD), apply Intel's stricter rules... */
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:55:10 +00:00
|
|
|
static int em_syscall(struct x86_emulate_ctxt *ctxt)
|
2009-06-18 10:56:00 +00:00
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-04-28 16:15:30 +00:00
|
|
|
struct desc_struct cs, ss;
|
2009-06-18 10:56:00 +00:00
|
|
|
u64 msr_data;
|
2010-04-28 16:15:30 +00:00
|
|
|
u16 cs_sel, ss_sel;
|
2011-04-20 12:21:35 +00:00
|
|
|
u64 efer = 0;
|
2009-06-18 10:56:00 +00:00
|
|
|
|
|
|
|
|
/* syscall is not available in real mode */
|
2010-03-18 13:20:12 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_REAL ||
|
2010-11-22 15:53:25 +00:00
|
|
|
ctxt->mode == X86EMUL_MODE_VM86)
|
|
|
|
|
return emulate_ud(ctxt);
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2012-01-12 15:43:04 +00:00
|
|
|
if (!(em_syscall_is_enabled(ctxt)))
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
2011-04-20 12:21:35 +00:00
|
|
|
ops->get_msr(ctxt, MSR_EFER, &efer);
|
2011-05-14 16:00:52 +00:00
|
|
|
setup_syscalls_segments(ctxt, &cs, &ss);
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2012-01-12 15:43:04 +00:00
|
|
|
if (!(efer & EFER_SCE))
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_STAR, &msr_data);
|
2009-06-18 10:56:00 +00:00
|
|
|
msr_data >>= 32;
|
2010-04-28 16:15:30 +00:00
|
|
|
cs_sel = (u16)(msr_data & 0xfffc);
|
|
|
|
|
ss_sel = (u16)(msr_data + 8);
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2011-04-20 12:21:35 +00:00
|
|
|
if (efer & EFER_LMA) {
|
2010-04-28 16:15:30 +00:00
|
|
|
cs.d = 0;
|
2009-06-18 10:56:00 +00:00
|
|
|
cs.l = 1;
|
|
|
|
|
}
|
2011-04-27 10:20:30 +00:00
|
|
|
ops->set_segment(ctxt, cs_sel, &cs, 0, VCPU_SREG_CS);
|
|
|
|
|
ops->set_segment(ctxt, ss_sel, &ss, 0, VCPU_SREG_SS);
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RCX) = ctxt->_eip;
|
2011-04-20 12:21:35 +00:00
|
|
|
if (efer & EFER_LMA) {
|
2009-06-18 10:56:00 +00:00
|
|
|
#ifdef CONFIG_X86_64
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_R11) = ctxt->eflags & ~EFLG_RF;
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt,
|
2010-04-28 16:15:28 +00:00
|
|
|
ctxt->mode == X86EMUL_MODE_PROT64 ?
|
|
|
|
|
MSR_LSTAR : MSR_CSTAR, &msr_data);
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = msr_data;
|
2009-06-18 10:56:00 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_SYSCALL_MASK, &msr_data);
|
2009-06-18 10:56:00 +00:00
|
|
|
ctxt->eflags &= ~(msr_data | EFLG_RF);
|
|
|
|
|
#endif
|
|
|
|
|
} else {
|
|
|
|
|
/* legacy mode */
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_STAR, &msr_data);
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = (u32)msr_data;
|
2009-06-18 10:56:00 +00:00
|
|
|
|
|
|
|
|
ctxt->eflags &= ~(EFLG_VM | EFLG_IF | EFLG_RF);
|
|
|
|
|
}
|
|
|
|
|
|
2010-02-18 10:15:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2009-06-18 10:56:00 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:55:10 +00:00
|
|
|
static int em_sysenter(struct x86_emulate_ctxt *ctxt)
|
2009-06-18 10:56:01 +00:00
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-04-28 16:15:30 +00:00
|
|
|
struct desc_struct cs, ss;
|
2009-06-18 10:56:01 +00:00
|
|
|
u64 msr_data;
|
2010-04-28 16:15:30 +00:00
|
|
|
u16 cs_sel, ss_sel;
|
2011-04-20 12:21:35 +00:00
|
|
|
u64 efer = 0;
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
ops->get_msr(ctxt, MSR_EFER, &efer);
|
2010-02-10 12:21:31 +00:00
|
|
|
/* inject #GP if in real mode */
|
2010-11-22 15:53:25 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_REAL)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2012-02-01 10:23:21 +00:00
|
|
|
/*
|
|
|
|
|
* Not recognized on AMD in compat mode (but is recognized in legacy
|
|
|
|
|
* mode).
|
|
|
|
|
*/
|
|
|
|
|
if ((ctxt->mode == X86EMUL_MODE_PROT32) && (efer & EFER_LMA)
|
|
|
|
|
&& !vendor_intel(ctxt))
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
2009-06-18 10:56:01 +00:00
|
|
|
/* XXX sysenter/sysexit have not been tested in 64bit mode.
|
|
|
|
|
* Therefore, we inject an #UD.
|
|
|
|
|
*/
|
2010-11-22 15:53:25 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
return emulate_ud(ctxt);
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
setup_syscalls_segments(ctxt, &cs, &ss);
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_IA32_SYSENTER_CS, &msr_data);
|
2009-06-18 10:56:01 +00:00
|
|
|
switch (ctxt->mode) {
|
|
|
|
|
case X86EMUL_MODE_PROT32:
|
2010-11-22 15:53:25 +00:00
|
|
|
if ((msr_data & 0xfffc) == 0x0)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2009-06-18 10:56:01 +00:00
|
|
|
break;
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
2010-11-22 15:53:25 +00:00
|
|
|
if (msr_data == 0x0)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2009-06-18 10:56:01 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ctxt->eflags &= ~(EFLG_VM | EFLG_IF | EFLG_RF);
|
2010-04-28 16:15:30 +00:00
|
|
|
cs_sel = (u16)msr_data;
|
|
|
|
|
cs_sel &= ~SELECTOR_RPL_MASK;
|
|
|
|
|
ss_sel = cs_sel + 8;
|
|
|
|
|
ss_sel &= ~SELECTOR_RPL_MASK;
|
2011-04-20 12:21:35 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64 || (efer & EFER_LMA)) {
|
2010-04-28 16:15:30 +00:00
|
|
|
cs.d = 0;
|
2009-06-18 10:56:01 +00:00
|
|
|
cs.l = 1;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
ops->set_segment(ctxt, cs_sel, &cs, 0, VCPU_SREG_CS);
|
|
|
|
|
ops->set_segment(ctxt, ss_sel, &ss, 0, VCPU_SREG_SS);
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_IA32_SYSENTER_EIP, &msr_data);
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = msr_data;
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_IA32_SYSENTER_ESP, &msr_data);
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RSP) = msr_data;
|
2009-06-18 10:56:01 +00:00
|
|
|
|
2010-02-18 10:15:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2009-06-18 10:56:01 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:55:10 +00:00
|
|
|
static int em_sysexit(struct x86_emulate_ctxt *ctxt)
|
2009-06-18 10:56:02 +00:00
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-04-28 16:15:30 +00:00
|
|
|
struct desc_struct cs, ss;
|
2009-06-18 10:56:02 +00:00
|
|
|
u64 msr_data;
|
|
|
|
|
int usermode;
|
2011-05-15 15:25:10 +00:00
|
|
|
u16 cs_sel = 0, ss_sel = 0;
|
2009-06-18 10:56:02 +00:00
|
|
|
|
2010-02-10 12:21:31 +00:00
|
|
|
/* inject #GP if in real mode or Virtual 8086 mode */
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_REAL ||
|
2010-11-22 15:53:25 +00:00
|
|
|
ctxt->mode == X86EMUL_MODE_VM86)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2009-06-18 10:56:02 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
setup_syscalls_segments(ctxt, &cs, &ss);
|
2009-06-18 10:56:02 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->rex_prefix & 0x8) != 0x0)
|
2009-06-18 10:56:02 +00:00
|
|
|
usermode = X86EMUL_MODE_PROT64;
|
|
|
|
|
else
|
|
|
|
|
usermode = X86EMUL_MODE_PROT32;
|
|
|
|
|
|
|
|
|
|
cs.dpl = 3;
|
|
|
|
|
ss.dpl = 3;
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_msr(ctxt, MSR_IA32_SYSENTER_CS, &msr_data);
|
2009-06-18 10:56:02 +00:00
|
|
|
switch (usermode) {
|
|
|
|
|
case X86EMUL_MODE_PROT32:
|
2010-04-28 16:15:30 +00:00
|
|
|
cs_sel = (u16)(msr_data + 16);
|
2010-11-22 15:53:25 +00:00
|
|
|
if ((msr_data & 0xfffc) == 0x0)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2010-04-28 16:15:30 +00:00
|
|
|
ss_sel = (u16)(msr_data + 24);
|
2009-06-18 10:56:02 +00:00
|
|
|
break;
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
2010-04-28 16:15:30 +00:00
|
|
|
cs_sel = (u16)(msr_data + 32);
|
2010-11-22 15:53:25 +00:00
|
|
|
if (msr_data == 0x0)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
2010-04-28 16:15:30 +00:00
|
|
|
ss_sel = cs_sel + 8;
|
|
|
|
|
cs.d = 0;
|
2009-06-18 10:56:02 +00:00
|
|
|
cs.l = 1;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2010-04-28 16:15:30 +00:00
|
|
|
cs_sel |= SELECTOR_RPL_MASK;
|
|
|
|
|
ss_sel |= SELECTOR_RPL_MASK;
|
2009-06-18 10:56:02 +00:00
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
ops->set_segment(ctxt, cs_sel, &cs, 0, VCPU_SREG_CS);
|
|
|
|
|
ops->set_segment(ctxt, ss_sel, &ss, 0, VCPU_SREG_SS);
|
2009-06-18 10:56:02 +00:00
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
ctxt->_eip = reg_read(ctxt, VCPU_REGS_RDX);
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RSP) = reg_read(ctxt, VCPU_REGS_RCX);
|
2009-06-18 10:56:02 +00:00
|
|
|
|
2010-02-18 10:15:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
2009-06-18 10:56:02 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
static bool emulator_bad_iopl(struct x86_emulate_ctxt *ctxt)
|
2010-02-10 12:21:33 +00:00
|
|
|
{
|
|
|
|
|
int iopl;
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_REAL)
|
|
|
|
|
return false;
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_VM86)
|
|
|
|
|
return true;
|
|
|
|
|
iopl = (ctxt->eflags & X86_EFLAGS_IOPL) >> IOPL_SHIFT;
|
2011-05-14 16:00:52 +00:00
|
|
|
return ctxt->ops->cpl(ctxt) > iopl;
|
2010-02-10 12:21:33 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool emulator_io_port_access_allowed(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 port, u16 len)
|
|
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-04-28 16:15:30 +00:00
|
|
|
struct desc_struct tr_seg;
|
2011-03-07 12:55:06 +00:00
|
|
|
u32 base3;
|
2010-02-10 12:21:33 +00:00
|
|
|
int r;
|
2011-04-27 10:20:30 +00:00
|
|
|
u16 tr, io_bitmap_ptr, perm, bit_idx = port & 0x7;
|
2010-02-10 12:21:33 +00:00
|
|
|
unsigned mask = (1 << len) - 1;
|
2011-03-07 12:55:06 +00:00
|
|
|
unsigned long base;
|
2010-02-10 12:21:33 +00:00
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
ops->get_segment(ctxt, &tr, &tr_seg, &base3, VCPU_SREG_TR);
|
2010-04-28 16:15:30 +00:00
|
|
|
if (!tr_seg.p)
|
2010-02-10 12:21:33 +00:00
|
|
|
return false;
|
2010-04-28 16:15:30 +00:00
|
|
|
if (desc_limit_scaled(&tr_seg) < 103)
|
2010-02-10 12:21:33 +00:00
|
|
|
return false;
|
2011-03-07 12:55:06 +00:00
|
|
|
base = get_desc_base(&tr_seg);
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
base |= ((u64)base3) << 32;
|
|
|
|
|
#endif
|
2011-04-20 10:37:53 +00:00
|
|
|
r = ops->read_std(ctxt, base + 102, &io_bitmap_ptr, 2, NULL);
|
2010-02-10 12:21:33 +00:00
|
|
|
if (r != X86EMUL_CONTINUE)
|
|
|
|
|
return false;
|
2010-04-28 16:15:30 +00:00
|
|
|
if (io_bitmap_ptr + port/8 > desc_limit_scaled(&tr_seg))
|
2010-02-10 12:21:33 +00:00
|
|
|
return false;
|
2011-04-20 10:37:53 +00:00
|
|
|
r = ops->read_std(ctxt, base + io_bitmap_ptr + port/8, &perm, 2, NULL);
|
2010-02-10 12:21:33 +00:00
|
|
|
if (r != X86EMUL_CONTINUE)
|
|
|
|
|
return false;
|
|
|
|
|
if ((perm >> bit_idx) & mask)
|
|
|
|
|
return false;
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool emulator_io_permited(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 port, u16 len)
|
|
|
|
|
{
|
2010-08-02 09:47:51 +00:00
|
|
|
if (ctxt->perm_ok)
|
|
|
|
|
return true;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
if (emulator_bad_iopl(ctxt))
|
|
|
|
|
if (!emulator_io_port_access_allowed(ctxt, port, len))
|
2010-02-10 12:21:33 +00:00
|
|
|
return false;
|
2010-08-02 09:47:51 +00:00
|
|
|
|
|
|
|
|
ctxt->perm_ok = true;
|
|
|
|
|
|
2010-02-10 12:21:33 +00:00
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2010-03-18 13:20:17 +00:00
|
|
|
static void save_state_to_tss16(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct tss_segment_16 *tss)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
tss->ip = ctxt->_eip;
|
2010-03-18 13:20:17 +00:00
|
|
|
tss->flag = ctxt->eflags;
|
2012-08-27 20:46:17 +00:00
|
|
|
tss->ax = reg_read(ctxt, VCPU_REGS_RAX);
|
|
|
|
|
tss->cx = reg_read(ctxt, VCPU_REGS_RCX);
|
|
|
|
|
tss->dx = reg_read(ctxt, VCPU_REGS_RDX);
|
|
|
|
|
tss->bx = reg_read(ctxt, VCPU_REGS_RBX);
|
|
|
|
|
tss->sp = reg_read(ctxt, VCPU_REGS_RSP);
|
|
|
|
|
tss->bp = reg_read(ctxt, VCPU_REGS_RBP);
|
|
|
|
|
tss->si = reg_read(ctxt, VCPU_REGS_RSI);
|
|
|
|
|
tss->di = reg_read(ctxt, VCPU_REGS_RDI);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
tss->es = get_segment_selector(ctxt, VCPU_SREG_ES);
|
|
|
|
|
tss->cs = get_segment_selector(ctxt, VCPU_SREG_CS);
|
|
|
|
|
tss->ss = get_segment_selector(ctxt, VCPU_SREG_SS);
|
|
|
|
|
tss->ds = get_segment_selector(ctxt, VCPU_SREG_DS);
|
|
|
|
|
tss->ldt = get_segment_selector(ctxt, VCPU_SREG_LDTR);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int load_state_from_tss16(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct tss_segment_16 *tss)
|
|
|
|
|
{
|
|
|
|
|
int ret;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = tss->ip;
|
2010-03-18 13:20:17 +00:00
|
|
|
ctxt->eflags = tss->flag | 2;
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = tss->ax;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RCX) = tss->cx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = tss->dx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RBX) = tss->bx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RSP) = tss->sp;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RBP) = tss->bp;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RSI) = tss->si;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDI) = tss->di;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* SDM says that segment selectors are loaded before segment
|
|
|
|
|
* descriptors
|
|
|
|
|
*/
|
2011-04-27 10:20:30 +00:00
|
|
|
set_segment_selector(ctxt, tss->ldt, VCPU_SREG_LDTR);
|
|
|
|
|
set_segment_selector(ctxt, tss->es, VCPU_SREG_ES);
|
|
|
|
|
set_segment_selector(ctxt, tss->cs, VCPU_SREG_CS);
|
|
|
|
|
set_segment_selector(ctxt, tss->ss, VCPU_SREG_SS);
|
|
|
|
|
set_segment_selector(ctxt, tss->ds, VCPU_SREG_DS);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
|
|
|
|
/*
|
2012-06-28 07:19:51 +00:00
|
|
|
* Now load segment descriptors. If fault happens at this stage
|
2010-03-18 13:20:17 +00:00
|
|
|
* it is handled in a context of new task
|
|
|
|
|
*/
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->ldt, VCPU_SREG_LDTR);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->es, VCPU_SREG_ES);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->cs, VCPU_SREG_CS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->ss, VCPU_SREG_SS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->ds, VCPU_SREG_DS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int task_switch_16(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 tss_selector, u16 old_tss_sel,
|
|
|
|
|
ulong old_tss_base, struct desc_struct *new_desc)
|
|
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-03-18 13:20:17 +00:00
|
|
|
struct tss_segment_16 tss_seg;
|
|
|
|
|
int ret;
|
2010-11-22 15:53:22 +00:00
|
|
|
u32 new_tss_base = get_desc_base(new_desc);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->read_std(ctxt, old_tss_base, &tss_seg, sizeof tss_seg,
|
2010-11-22 15:53:22 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
save_state_to_tss16(ctxt, &tss_seg);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->write_std(ctxt, old_tss_base, &tss_seg, sizeof tss_seg,
|
2010-11-22 15:53:22 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->read_std(ctxt, new_tss_base, &tss_seg, sizeof tss_seg,
|
2010-11-22 15:53:22 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
if (old_tss_sel != 0xffff) {
|
|
|
|
|
tss_seg.prev_task_link = old_tss_sel;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->write_std(ctxt, new_tss_base,
|
2010-03-18 13:20:17 +00:00
|
|
|
&tss_seg.prev_task_link,
|
|
|
|
|
sizeof tss_seg.prev_task_link,
|
2011-04-20 10:37:53 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
return load_state_from_tss16(ctxt, &tss_seg);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void save_state_to_tss32(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct tss_segment_32 *tss)
|
|
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
tss->cr3 = ctxt->ops->get_cr(ctxt, 3);
|
2011-06-01 12:34:25 +00:00
|
|
|
tss->eip = ctxt->_eip;
|
2010-03-18 13:20:17 +00:00
|
|
|
tss->eflags = ctxt->eflags;
|
2012-08-27 20:46:17 +00:00
|
|
|
tss->eax = reg_read(ctxt, VCPU_REGS_RAX);
|
|
|
|
|
tss->ecx = reg_read(ctxt, VCPU_REGS_RCX);
|
|
|
|
|
tss->edx = reg_read(ctxt, VCPU_REGS_RDX);
|
|
|
|
|
tss->ebx = reg_read(ctxt, VCPU_REGS_RBX);
|
|
|
|
|
tss->esp = reg_read(ctxt, VCPU_REGS_RSP);
|
|
|
|
|
tss->ebp = reg_read(ctxt, VCPU_REGS_RBP);
|
|
|
|
|
tss->esi = reg_read(ctxt, VCPU_REGS_RSI);
|
|
|
|
|
tss->edi = reg_read(ctxt, VCPU_REGS_RDI);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
tss->es = get_segment_selector(ctxt, VCPU_SREG_ES);
|
|
|
|
|
tss->cs = get_segment_selector(ctxt, VCPU_SREG_CS);
|
|
|
|
|
tss->ss = get_segment_selector(ctxt, VCPU_SREG_SS);
|
|
|
|
|
tss->ds = get_segment_selector(ctxt, VCPU_SREG_DS);
|
|
|
|
|
tss->fs = get_segment_selector(ctxt, VCPU_SREG_FS);
|
|
|
|
|
tss->gs = get_segment_selector(ctxt, VCPU_SREG_GS);
|
|
|
|
|
tss->ldt_selector = get_segment_selector(ctxt, VCPU_SREG_LDTR);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int load_state_from_tss32(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct tss_segment_32 *tss)
|
|
|
|
|
{
|
|
|
|
|
int ret;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
if (ctxt->ops->set_cr(ctxt, 3, tss->cr3))
|
2010-11-22 15:53:25 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = tss->eip;
|
2010-03-18 13:20:17 +00:00
|
|
|
ctxt->eflags = tss->eflags | 2;
|
2012-02-08 13:34:41 +00:00
|
|
|
|
|
|
|
|
/* General purpose registers */
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = tss->eax;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RCX) = tss->ecx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = tss->edx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RBX) = tss->ebx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RSP) = tss->esp;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RBP) = tss->ebp;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RSI) = tss->esi;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDI) = tss->edi;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* SDM says that segment selectors are loaded before segment
|
|
|
|
|
* descriptors
|
|
|
|
|
*/
|
2011-04-27 10:20:30 +00:00
|
|
|
set_segment_selector(ctxt, tss->ldt_selector, VCPU_SREG_LDTR);
|
|
|
|
|
set_segment_selector(ctxt, tss->es, VCPU_SREG_ES);
|
|
|
|
|
set_segment_selector(ctxt, tss->cs, VCPU_SREG_CS);
|
|
|
|
|
set_segment_selector(ctxt, tss->ss, VCPU_SREG_SS);
|
|
|
|
|
set_segment_selector(ctxt, tss->ds, VCPU_SREG_DS);
|
|
|
|
|
set_segment_selector(ctxt, tss->fs, VCPU_SREG_FS);
|
|
|
|
|
set_segment_selector(ctxt, tss->gs, VCPU_SREG_GS);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2012-02-08 13:34:41 +00:00
|
|
|
/*
|
|
|
|
|
* If we're switching between Protected Mode and VM86, we need to make
|
|
|
|
|
* sure to update the mode before loading the segment descriptors so
|
|
|
|
|
* that the selectors are interpreted correctly.
|
|
|
|
|
*
|
|
|
|
|
* Need to get rflags to the vcpu struct immediately because it
|
|
|
|
|
* influences the CPL which is checked at least when loading the segment
|
|
|
|
|
* descriptors and when pushing an error code to the new kernel stack.
|
|
|
|
|
*
|
|
|
|
|
* TODO Introduce a separate ctxt->ops->set_cpl callback
|
|
|
|
|
*/
|
|
|
|
|
if (ctxt->eflags & X86_EFLAGS_VM)
|
|
|
|
|
ctxt->mode = X86EMUL_MODE_VM86;
|
|
|
|
|
else
|
|
|
|
|
ctxt->mode = X86EMUL_MODE_PROT32;
|
|
|
|
|
|
|
|
|
|
ctxt->ops->set_rflags(ctxt, ctxt->eflags);
|
|
|
|
|
|
2010-03-18 13:20:17 +00:00
|
|
|
/*
|
|
|
|
|
* Now load segment descriptors. If fault happenes at this stage
|
|
|
|
|
* it is handled in a context of new task
|
|
|
|
|
*/
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->ldt_selector, VCPU_SREG_LDTR);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->es, VCPU_SREG_ES);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->cs, VCPU_SREG_CS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->ss, VCPU_SREG_SS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->ds, VCPU_SREG_DS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->fs, VCPU_SREG_FS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = load_segment_descriptor(ctxt, tss->gs, VCPU_SREG_GS);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int task_switch_32(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
u16 tss_selector, u16 old_tss_sel,
|
|
|
|
|
ulong old_tss_base, struct desc_struct *new_desc)
|
|
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-03-18 13:20:17 +00:00
|
|
|
struct tss_segment_32 tss_seg;
|
|
|
|
|
int ret;
|
2010-11-22 15:53:22 +00:00
|
|
|
u32 new_tss_base = get_desc_base(new_desc);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->read_std(ctxt, old_tss_base, &tss_seg, sizeof tss_seg,
|
2010-11-22 15:53:22 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
save_state_to_tss32(ctxt, &tss_seg);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->write_std(ctxt, old_tss_base, &tss_seg, sizeof tss_seg,
|
2010-11-22 15:53:22 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->read_std(ctxt, new_tss_base, &tss_seg, sizeof tss_seg,
|
2010-11-22 15:53:22 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
if (old_tss_sel != 0xffff) {
|
|
|
|
|
tss_seg.prev_task_link = old_tss_sel;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ret = ops->write_std(ctxt, new_tss_base,
|
2010-03-18 13:20:17 +00:00
|
|
|
&tss_seg.prev_task_link,
|
|
|
|
|
sizeof tss_seg.prev_task_link,
|
2011-04-20 10:37:53 +00:00
|
|
|
&ctxt->exception);
|
2010-11-22 15:53:24 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
2010-03-18 13:20:17 +00:00
|
|
|
/* FIXME: need to provide precise fault address */
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
return load_state_from_tss32(ctxt, &tss_seg);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int emulator_do_task_switch(struct x86_emulate_ctxt *ctxt,
|
2012-02-08 13:34:38 +00:00
|
|
|
u16 tss_selector, int idt_index, int reason,
|
2010-04-14 13:51:09 +00:00
|
|
|
bool has_error_code, u32 error_code)
|
2010-03-18 13:20:17 +00:00
|
|
|
{
|
2011-05-14 16:00:52 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-03-18 13:20:17 +00:00
|
|
|
struct desc_struct curr_tss_desc, next_tss_desc;
|
|
|
|
|
int ret;
|
2011-04-27 10:20:30 +00:00
|
|
|
u16 old_tss_sel = get_segment_selector(ctxt, VCPU_SREG_TR);
|
2010-03-18 13:20:17 +00:00
|
|
|
ulong old_tss_base =
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->get_cached_segment_base(ctxt, VCPU_SREG_TR);
|
2010-03-18 13:20:19 +00:00
|
|
|
u32 desc_limit;
|
2012-06-13 13:29:39 +00:00
|
|
|
ulong desc_addr;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
|
|
|
|
/* FIXME: old_tss_base == ~0 ? */
|
|
|
|
|
|
2012-06-13 13:29:39 +00:00
|
|
|
ret = read_segment_descriptor(ctxt, tss_selector, &next_tss_desc, &desc_addr);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2012-06-13 13:29:39 +00:00
|
|
|
ret = read_segment_descriptor(ctxt, old_tss_sel, &curr_tss_desc, &desc_addr);
|
2010-03-18 13:20:17 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
/* FIXME: check that next_tss_desc is tss */
|
|
|
|
|
|
2012-02-08 13:34:38 +00:00
|
|
|
/*
|
|
|
|
|
* Check privileges. The three cases are task switch caused by...
|
|
|
|
|
*
|
|
|
|
|
* 1. jmp/call/int to task gate: Check against DPL of the task gate
|
|
|
|
|
* 2. Exception/IRQ/iret: No check is performed
|
2012-06-28 07:19:51 +00:00
|
|
|
* 3. jmp/call to TSS: Check against DPL of the TSS
|
2012-02-08 13:34:38 +00:00
|
|
|
*/
|
|
|
|
|
if (reason == TASK_SWITCH_GATE) {
|
|
|
|
|
if (idt_index != -1) {
|
|
|
|
|
/* Software interrupts */
|
|
|
|
|
struct desc_struct task_gate_desc;
|
|
|
|
|
int dpl;
|
|
|
|
|
|
|
|
|
|
ret = read_interrupt_descriptor(ctxt, idt_index,
|
|
|
|
|
&task_gate_desc);
|
|
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
|
|
dpl = task_gate_desc.dpl;
|
|
|
|
|
if ((tss_selector & 3) > dpl || ops->cpl(ctxt) > dpl)
|
|
|
|
|
return emulate_gp(ctxt, (idt_index << 3) | 0x2);
|
|
|
|
|
}
|
|
|
|
|
} else if (reason != TASK_SWITCH_IRET) {
|
|
|
|
|
int dpl = next_tss_desc.dpl;
|
|
|
|
|
if ((tss_selector & 3) > dpl || ops->cpl(ctxt) > dpl)
|
|
|
|
|
return emulate_gp(ctxt, tss_selector);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2012-02-08 13:34:38 +00:00
|
|
|
|
2010-03-18 13:20:19 +00:00
|
|
|
desc_limit = desc_limit_scaled(&next_tss_desc);
|
|
|
|
|
if (!next_tss_desc.p ||
|
|
|
|
|
((desc_limit < 0x67 && (next_tss_desc.type & 8)) ||
|
|
|
|
|
desc_limit < 0x2b)) {
|
2010-04-28 16:15:44 +00:00
|
|
|
emulate_ts(ctxt, tss_selector & 0xfffc);
|
2010-03-18 13:20:17 +00:00
|
|
|
return X86EMUL_PROPAGATE_FAULT;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (reason == TASK_SWITCH_IRET || reason == TASK_SWITCH_JMP) {
|
|
|
|
|
curr_tss_desc.type &= ~(1 << 1); /* clear busy flag */
|
2011-05-14 16:00:52 +00:00
|
|
|
write_segment_descriptor(ctxt, old_tss_sel, &curr_tss_desc);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (reason == TASK_SWITCH_IRET)
|
|
|
|
|
ctxt->eflags = ctxt->eflags & ~X86_EFLAGS_NT;
|
|
|
|
|
|
|
|
|
|
/* set back link to prev task only if NT bit is set in eflags
|
2012-06-28 07:19:51 +00:00
|
|
|
note that old_tss_sel is not used after this point */
|
2010-03-18 13:20:17 +00:00
|
|
|
if (reason != TASK_SWITCH_CALL && reason != TASK_SWITCH_GATE)
|
|
|
|
|
old_tss_sel = 0xffff;
|
|
|
|
|
|
|
|
|
|
if (next_tss_desc.type & 8)
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = task_switch_32(ctxt, tss_selector, old_tss_sel,
|
2010-03-18 13:20:17 +00:00
|
|
|
old_tss_base, &next_tss_desc);
|
|
|
|
|
else
|
2011-05-14 16:00:52 +00:00
|
|
|
ret = task_switch_16(ctxt, tss_selector, old_tss_sel,
|
2010-03-18 13:20:17 +00:00
|
|
|
old_tss_base, &next_tss_desc);
|
2010-04-14 13:50:57 +00:00
|
|
|
if (ret != X86EMUL_CONTINUE)
|
|
|
|
|
return ret;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
|
|
|
|
if (reason == TASK_SWITCH_CALL || reason == TASK_SWITCH_GATE)
|
|
|
|
|
ctxt->eflags = ctxt->eflags | X86_EFLAGS_NT;
|
|
|
|
|
|
|
|
|
|
if (reason != TASK_SWITCH_IRET) {
|
|
|
|
|
next_tss_desc.type |= (1 << 1); /* set busy flag */
|
2011-05-14 16:00:52 +00:00
|
|
|
write_segment_descriptor(ctxt, tss_selector, &next_tss_desc);
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ops->set_cr(ctxt, 0, ops->get_cr(ctxt, 0) | X86_CR0_TS);
|
2011-04-27 10:20:30 +00:00
|
|
|
ops->set_segment(ctxt, tss_selector, &next_tss_desc, 0, VCPU_SREG_TR);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2010-04-14 13:51:09 +00:00
|
|
|
if (has_error_code) {
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes = ctxt->ad_bytes = (next_tss_desc.type & 8) ? 4 : 2;
|
|
|
|
|
ctxt->lock_prefix = 0;
|
|
|
|
|
ctxt->src.val = (unsigned long) error_code;
|
2011-04-12 15:31:23 +00:00
|
|
|
ret = em_push(ctxt);
|
2010-04-14 13:51:09 +00:00
|
|
|
}
|
|
|
|
|
|
2010-03-18 13:20:17 +00:00
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int emulator_task_switch(struct x86_emulate_ctxt *ctxt,
|
2012-02-08 13:34:38 +00:00
|
|
|
u16 tss_selector, int idt_index, int reason,
|
2010-04-14 13:51:09 +00:00
|
|
|
bool has_error_code, u32 error_code)
|
2010-03-18 13:20:17 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
invalidate_registers(ctxt);
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = ctxt->eip;
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2012-02-08 13:34:38 +00:00
|
|
|
rc = emulator_do_task_switch(ctxt, tss_selector, idt_index, reason,
|
2010-04-14 13:51:09 +00:00
|
|
|
has_error_code, error_code);
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
if (rc == X86EMUL_CONTINUE) {
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->eip = ctxt->_eip;
|
2012-08-27 20:46:17 +00:00
|
|
|
writeback_registers(ctxt);
|
|
|
|
|
}
|
2010-03-18 13:20:17 +00:00
|
|
|
|
2011-03-28 14:57:49 +00:00
|
|
|
return (rc == X86EMUL_UNHANDLEABLE) ? EMULATION_FAILED : EMULATION_OK;
|
2010-03-18 13:20:17 +00:00
|
|
|
}
|
|
|
|
|
|
2010-11-17 13:28:21 +00:00
|
|
|
static void string_addr_inc(struct x86_emulate_ctxt *ctxt, unsigned seg,
|
2010-03-18 13:20:22 +00:00
|
|
|
int reg, struct operand *op)
|
2010-03-18 13:20:21 +00:00
|
|
|
{
|
|
|
|
|
int df = (ctxt->eflags & EFLG_DF) ? -1 : 1;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
register_address_increment(ctxt, reg_rmw(ctxt, reg), df * op->bytes);
|
|
|
|
|
op->addr.mem.ea = register_address(ctxt, reg_read(ctxt, reg));
|
2010-11-17 13:28:21 +00:00
|
|
|
op->addr.mem.seg = seg;
|
2010-03-18 13:20:21 +00:00
|
|
|
}
|
|
|
|
|
|
2010-08-18 11:16:35 +00:00
|
|
|
static int em_das(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u8 al, old_al;
|
|
|
|
|
bool af, cf, old_cf;
|
|
|
|
|
|
|
|
|
|
cf = ctxt->eflags & X86_EFLAGS_CF;
|
2011-06-01 12:34:25 +00:00
|
|
|
al = ctxt->dst.val;
|
2010-08-18 11:16:35 +00:00
|
|
|
|
|
|
|
|
old_al = al;
|
|
|
|
|
old_cf = cf;
|
|
|
|
|
cf = false;
|
|
|
|
|
af = ctxt->eflags & X86_EFLAGS_AF;
|
|
|
|
|
if ((al & 0x0f) > 9 || af) {
|
|
|
|
|
al -= 6;
|
|
|
|
|
cf = old_cf | (al >= 250);
|
|
|
|
|
af = true;
|
|
|
|
|
} else {
|
|
|
|
|
af = false;
|
|
|
|
|
}
|
|
|
|
|
if (old_al > 0x99 || old_cf) {
|
|
|
|
|
al -= 0x60;
|
|
|
|
|
cf = true;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = al;
|
2010-08-18 11:16:35 +00:00
|
|
|
/* Set PF, ZF, SF */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.type = OP_IMM;
|
|
|
|
|
ctxt->src.val = 0;
|
|
|
|
|
ctxt->src.bytes = 1;
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "or");
|
2010-08-18 11:16:35 +00:00
|
|
|
ctxt->eflags &= ~(X86_EFLAGS_AF | X86_EFLAGS_CF);
|
|
|
|
|
if (cf)
|
|
|
|
|
ctxt->eflags |= X86_EFLAGS_CF;
|
|
|
|
|
if (af)
|
|
|
|
|
ctxt->eflags |= X86_EFLAGS_AF;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:18:35 +00:00
|
|
|
static int em_call(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
long rel = ctxt->src.val;
|
|
|
|
|
|
|
|
|
|
ctxt->src.val = (unsigned long)ctxt->_eip;
|
|
|
|
|
jmp_rel(ctxt, rel);
|
|
|
|
|
return em_push(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-18 11:51:45 +00:00
|
|
|
static int em_call_far(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u16 sel, old_cs;
|
|
|
|
|
ulong old_eip;
|
|
|
|
|
int rc;
|
|
|
|
|
|
2011-04-27 10:20:30 +00:00
|
|
|
old_cs = get_segment_selector(ctxt, VCPU_SREG_CS);
|
2011-06-01 12:34:25 +00:00
|
|
|
old_eip = ctxt->_eip;
|
2010-08-18 11:51:45 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
memcpy(&sel, ctxt->src.valptr + ctxt->op_bytes, 2);
|
2011-05-14 16:00:52 +00:00
|
|
|
if (load_segment_descriptor(ctxt, sel, VCPU_SREG_CS))
|
2010-08-18 11:51:45 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = 0;
|
|
|
|
|
memcpy(&ctxt->_eip, ctxt->src.valptr, ctxt->op_bytes);
|
2010-08-18 11:51:45 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = old_cs;
|
2011-04-12 15:31:23 +00:00
|
|
|
rc = em_push(ctxt);
|
2010-08-18 11:51:45 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = old_eip;
|
2011-04-12 15:31:23 +00:00
|
|
|
return em_push(ctxt);
|
2010-08-18 11:51:45 +00:00
|
|
|
}
|
|
|
|
|
|
2010-08-18 12:12:09 +00:00
|
|
|
static int em_ret_near_imm(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_REG;
|
|
|
|
|
ctxt->dst.addr.reg = &ctxt->_eip;
|
|
|
|
|
ctxt->dst.bytes = ctxt->op_bytes;
|
|
|
|
|
rc = emulate_pop(ctxt, &ctxt->dst.val, ctxt->op_bytes);
|
2010-08-18 12:12:09 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
2012-08-19 11:34:31 +00:00
|
|
|
rsp_increment(ctxt, ctxt->src.val);
|
2010-08-18 12:12:09 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
static int em_add(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "add");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_or(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "or");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_adc(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "adc");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_sbb(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "sbb");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_and(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "and");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_sub(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "sub");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_xor(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "xor");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_cmp(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "cmp");
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:57:53 +00:00
|
|
|
static int em_test(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "test");
|
2011-09-11 08:23:02 +00:00
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-05-29 12:57:53 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 12:59:09 +00:00
|
|
|
static int em_xchg(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
/* Write back the register source. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = ctxt->dst.val;
|
|
|
|
|
write_register_operand(&ctxt->src);
|
2011-05-29 12:59:09 +00:00
|
|
|
|
|
|
|
|
/* Write back the memory destination with implicit LOCK prefix. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ctxt->src.orig_val;
|
|
|
|
|
ctxt->lock_prefix = 1;
|
2011-05-29 12:59:09 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-18 15:31:43 +00:00
|
|
|
static int em_imul(struct x86_emulate_ctxt *ctxt)
|
2010-08-18 15:25:25 +00:00
|
|
|
{
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "imul");
|
2010-08-18 15:25:25 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-18 15:31:43 +00:00
|
|
|
static int em_imul_3op(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ctxt->src2.val;
|
2010-08-18 15:31:43 +00:00
|
|
|
return em_imul(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-19 12:13:00 +00:00
|
|
|
static int em_cwd(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_REG;
|
|
|
|
|
ctxt->dst.bytes = ctxt->src.bytes;
|
2012-08-27 20:46:17 +00:00
|
|
|
ctxt->dst.addr.reg = reg_rmw(ctxt, VCPU_REGS_RDX);
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ~((ctxt->src.val >> (ctxt->src.bytes * 8 - 1)) - 1);
|
2010-08-19 12:13:00 +00:00
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-18 15:54:34 +00:00
|
|
|
static int em_rdtsc(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u64 tsc = 0;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ctxt->ops->get_msr(ctxt, MSR_IA32_TSC, &tsc);
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = (u32)tsc;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = tsc >> 32;
|
2010-08-18 15:54:34 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-10 12:57:30 +00:00
|
|
|
static int em_rdpmc(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u64 pmc;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
if (ctxt->ops->read_pmc(ctxt, reg_read(ctxt, VCPU_REGS_RCX), &pmc))
|
2011-11-10 12:57:30 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = (u32)pmc;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = pmc >> 32;
|
2011-11-10 12:57:30 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-03 11:46:56 +00:00
|
|
|
static int em_mov(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-04-09 15:40:00 +00:00
|
|
|
memcpy(ctxt->dst.valptr, ctxt->src.valptr, ctxt->op_bytes);
|
2010-08-03 11:46:56 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:19:19 +00:00
|
|
|
static int em_cr_write(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
if (ctxt->ops->set_cr(ctxt, ctxt->modrm_reg, ctxt->src.val))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_dr_write(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
unsigned long val;
|
|
|
|
|
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
val = ctxt->src.val & ~0ULL;
|
|
|
|
|
else
|
|
|
|
|
val = ctxt->src.val & ~0U;
|
|
|
|
|
|
|
|
|
|
/* #UD condition is already handled. */
|
|
|
|
|
if (ctxt->ops->set_dr(ctxt, ctxt->modrm_reg, val) < 0)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:20:03 +00:00
|
|
|
static int em_wrmsr(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u64 msr_data;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
msr_data = (u32)reg_read(ctxt, VCPU_REGS_RAX)
|
|
|
|
|
| ((u64)reg_read(ctxt, VCPU_REGS_RDX) << 32);
|
|
|
|
|
if (ctxt->ops->set_msr(ctxt, reg_read(ctxt, VCPU_REGS_RCX), msr_data))
|
2011-11-22 06:20:03 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_rdmsr(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u64 msr_data;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
if (ctxt->ops->get_msr(ctxt, reg_read(ctxt, VCPU_REGS_RCX), &msr_data))
|
2011-11-22 06:20:03 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = (u32)msr_data;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = msr_data >> 32;
|
2011-11-22 06:20:03 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 13:01:33 +00:00
|
|
|
static int em_mov_rm_sreg(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_reg > VCPU_SREG_GS)
|
2011-05-29 13:01:33 +00:00
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = get_segment_selector(ctxt, ctxt->modrm_reg);
|
2011-05-29 13:01:33 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_mov_sreg_rm(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
u16 sel = ctxt->src.val;
|
2011-05-29 13:01:33 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_reg == VCPU_SREG_CS || ctxt->modrm_reg > VCPU_SREG_GS)
|
2011-05-29 13:01:33 +00:00
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_reg == VCPU_SREG_SS)
|
2011-05-29 13:01:33 +00:00
|
|
|
ctxt->interruptibility = KVM_X86_SHADOW_INT_MOV_SS;
|
|
|
|
|
|
|
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return load_segment_descriptor(ctxt, sel, ctxt->modrm_reg);
|
2011-05-29 13:01:33 +00:00
|
|
|
}
|
|
|
|
|
|
2012-06-13 09:28:33 +00:00
|
|
|
static int em_lldt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u16 sel = ctxt->src.val;
|
|
|
|
|
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return load_segment_descriptor(ctxt, sel, VCPU_SREG_LDTR);
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-13 13:33:29 +00:00
|
|
|
static int em_ltr(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u16 sel = ctxt->src.val;
|
|
|
|
|
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return load_segment_descriptor(ctxt, sel, VCPU_SREG_TR);
|
|
|
|
|
}
|
|
|
|
|
|
2011-03-31 16:48:09 +00:00
|
|
|
static int em_invlpg(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-03-31 16:54:30 +00:00
|
|
|
int rc;
|
|
|
|
|
ulong linear;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = linearize(ctxt, ctxt->src.addr.mem, 1, false, &linear);
|
2011-03-31 16:54:30 +00:00
|
|
|
if (rc == X86EMUL_CONTINUE)
|
2011-04-20 12:38:44 +00:00
|
|
|
ctxt->ops->invlpg(ctxt, linear);
|
2011-03-31 16:48:09 +00:00
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-03-31 16:48:09 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-20 12:32:49 +00:00
|
|
|
static int em_clts(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
ulong cr0;
|
|
|
|
|
|
|
|
|
|
cr0 = ctxt->ops->get_cr(ctxt, 0);
|
|
|
|
|
cr0 &= ~X86_CR0_TS;
|
|
|
|
|
ctxt->ops->set_cr(ctxt, 0, cr0);
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-21 09:07:59 +00:00
|
|
|
static int em_vmcall(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->modrm_mod != 3 || ctxt->modrm_rm != 1)
|
2011-04-21 09:07:59 +00:00
|
|
|
return X86EMUL_UNHANDLEABLE;
|
|
|
|
|
|
|
|
|
|
rc = ctxt->ops->fix_hypercall(ctxt);
|
|
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
|
|
|
|
|
/* Let the processor re-execute the fixed hypercall */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = ctxt->eip;
|
2011-04-21 09:07:59 +00:00
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-04-21 09:07:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-10 14:21:18 +00:00
|
|
|
static int emulate_store_desc_ptr(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
void (*get)(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct desc_ptr *ptr))
|
|
|
|
|
{
|
|
|
|
|
struct desc_ptr desc_ptr;
|
|
|
|
|
|
|
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
ctxt->op_bytes = 8;
|
|
|
|
|
get(ctxt, &desc_ptr);
|
|
|
|
|
if (ctxt->op_bytes == 2) {
|
|
|
|
|
ctxt->op_bytes = 4;
|
|
|
|
|
desc_ptr.address &= 0x00ffffff;
|
|
|
|
|
}
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return segmented_write(ctxt, ctxt->dst.addr.mem,
|
|
|
|
|
&desc_ptr, 2 + ctxt->op_bytes);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_sgdt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
return emulate_store_desc_ptr(ctxt, ctxt->ops->get_gdt);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_sidt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
return emulate_store_desc_ptr(ctxt, ctxt->ops->get_idt);
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-21 09:07:59 +00:00
|
|
|
static int em_lgdt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
struct desc_ptr desc_ptr;
|
|
|
|
|
int rc;
|
|
|
|
|
|
2012-06-07 14:04:36 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
ctxt->op_bytes = 8;
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = read_descriptor(ctxt, ctxt->src.addr.mem,
|
2011-04-21 09:07:59 +00:00
|
|
|
&desc_ptr.size, &desc_ptr.address,
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes);
|
2011-04-21 09:07:59 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
ctxt->ops->set_gdt(ctxt, &desc_ptr);
|
|
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-04-21 09:07:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-21 09:21:50 +00:00
|
|
|
static int em_vmmcall(struct x86_emulate_ctxt *ctxt)
|
2011-04-21 09:07:59 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
2011-04-21 09:21:50 +00:00
|
|
|
rc = ctxt->ops->fix_hypercall(ctxt);
|
|
|
|
|
|
2011-04-21 09:07:59 +00:00
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-04-21 09:07:59 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_lidt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
struct desc_ptr desc_ptr;
|
|
|
|
|
int rc;
|
|
|
|
|
|
2012-06-07 14:04:36 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
ctxt->op_bytes = 8;
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = read_descriptor(ctxt, ctxt->src.addr.mem,
|
2011-05-01 17:25:07 +00:00
|
|
|
&desc_ptr.size, &desc_ptr.address,
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes);
|
2011-04-21 09:07:59 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
return rc;
|
|
|
|
|
ctxt->ops->set_idt(ctxt, &desc_ptr);
|
|
|
|
|
/* Disable writeback. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-04-21 09:07:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_smsw(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.bytes = 2;
|
|
|
|
|
ctxt->dst.val = ctxt->ops->get_cr(ctxt, 0);
|
2011-04-21 09:07:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_lmsw(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
ctxt->ops->set_cr(ctxt, 0, (ctxt->ops->get_cr(ctxt, 0) & ~0x0eul)
|
2011-06-01 12:34:25 +00:00
|
|
|
| (ctxt->src.val & 0x0f));
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
2011-04-21 09:07:59 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 13:04:08 +00:00
|
|
|
static int em_loop(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
register_address_increment(ctxt, reg_rmw(ctxt, VCPU_REGS_RCX), -1);
|
|
|
|
|
if ((address_mask(ctxt, reg_read(ctxt, VCPU_REGS_RCX)) != 0) &&
|
2011-06-01 12:34:25 +00:00
|
|
|
(ctxt->b == 0xe2 || test_cc(ctxt->b ^ 0x5, ctxt->eflags)))
|
|
|
|
|
jmp_rel(ctxt, ctxt->src.val);
|
2011-05-29 13:04:08 +00:00
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_jcxz(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
if (address_mask(ctxt, reg_read(ctxt, VCPU_REGS_RCX)) == 0)
|
2011-06-01 12:34:25 +00:00
|
|
|
jmp_rel(ctxt, ctxt->src.val);
|
2011-05-29 13:04:08 +00:00
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:16:54 +00:00
|
|
|
static int em_in(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
if (!pio_in_emulated(ctxt, ctxt->dst.bytes, ctxt->src.val,
|
|
|
|
|
&ctxt->dst.val))
|
|
|
|
|
return X86EMUL_IO_NEEDED;
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_out(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
ctxt->ops->pio_out_emulated(ctxt, ctxt->src.bytes, ctxt->dst.val,
|
|
|
|
|
&ctxt->src.val, 1);
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-29 13:05:15 +00:00
|
|
|
static int em_cli(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
if (emulator_bad_iopl(ctxt))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
ctxt->eflags &= ~X86_EFLAGS_IF;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_sti(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
if (emulator_bad_iopl(ctxt))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
ctxt->interruptibility = KVM_X86_SHADOW_INT_STI;
|
|
|
|
|
ctxt->eflags |= X86_EFLAGS_IF;
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:17:48 +00:00
|
|
|
static int em_bt(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
/* Disable writeback. */
|
|
|
|
|
ctxt->dst.type = OP_NONE;
|
|
|
|
|
/* only subword offset */
|
|
|
|
|
ctxt->src.val &= (ctxt->dst.bytes << 3) - 1;
|
|
|
|
|
|
|
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "bt");
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_bts(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "bts");
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_btr(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "btr");
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_btc(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "btc");
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-11-22 06:21:33 +00:00
|
|
|
static int em_bsf(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-05-07 10:12:25 +00:00
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "bsf");
|
2011-11-22 06:21:33 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int em_bsr(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-05-07 10:12:25 +00:00
|
|
|
emulate_2op_SrcV_nobyte(ctxt, "bsr");
|
2011-11-22 06:21:33 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-07 11:11:36 +00:00
|
|
|
static int em_cpuid(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u32 eax, ebx, ecx, edx;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
eax = reg_read(ctxt, VCPU_REGS_RAX);
|
|
|
|
|
ecx = reg_read(ctxt, VCPU_REGS_RCX);
|
2012-06-07 11:11:36 +00:00
|
|
|
ctxt->ops->get_cpuid(ctxt, &eax, &ebx, &ecx, &edx);
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_write(ctxt, VCPU_REGS_RAX) = eax;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RBX) = ebx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RCX) = ecx;
|
|
|
|
|
*reg_write(ctxt, VCPU_REGS_RDX) = edx;
|
2012-06-07 11:11:36 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-11 10:09:07 +00:00
|
|
|
static int em_lahf(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
*reg_rmw(ctxt, VCPU_REGS_RAX) &= ~0xff00UL;
|
|
|
|
|
*reg_rmw(ctxt, VCPU_REGS_RAX) |= (ctxt->eflags & 0xff) << 8;
|
2012-06-11 10:09:07 +00:00
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2012-06-13 09:25:06 +00:00
|
|
|
static int em_bswap(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
switch (ctxt->op_bytes) {
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
case 8:
|
|
|
|
|
asm("bswap %0" : "+r"(ctxt->dst.val));
|
|
|
|
|
break;
|
|
|
|
|
#endif
|
|
|
|
|
default:
|
|
|
|
|
asm("bswap %0" : "+r"(*(u32 *)&ctxt->dst.val));
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:28 +00:00
|
|
|
static bool valid_cr(int nr)
|
|
|
|
|
{
|
|
|
|
|
switch (nr) {
|
|
|
|
|
case 0:
|
|
|
|
|
case 2 ... 4:
|
|
|
|
|
case 8:
|
|
|
|
|
return true;
|
|
|
|
|
default:
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int check_cr_read(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
if (!valid_cr(ctxt->modrm_reg))
|
2011-04-04 10:39:28 +00:00
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int check_cr_write(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
u64 new_val = ctxt->src.val64;
|
|
|
|
|
int cr = ctxt->modrm_reg;
|
2011-04-20 12:21:35 +00:00
|
|
|
u64 efer = 0;
|
2011-04-04 10:39:28 +00:00
|
|
|
|
|
|
|
|
static u64 cr_reserved_bits[] = {
|
|
|
|
|
0xffffffff00000000ULL,
|
|
|
|
|
0, 0, 0, /* CR3 checked later */
|
|
|
|
|
CR4_RESERVED_BITS,
|
|
|
|
|
0, 0, 0,
|
|
|
|
|
CR8_RESERVED_BITS,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
if (!valid_cr(cr))
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
|
|
|
|
if (new_val & cr_reserved_bits[cr])
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
switch (cr) {
|
|
|
|
|
case 0: {
|
2011-04-20 12:21:35 +00:00
|
|
|
u64 cr4;
|
2011-04-04 10:39:28 +00:00
|
|
|
if (((new_val & X86_CR0_PG) && !(new_val & X86_CR0_PE)) ||
|
|
|
|
|
((new_val & X86_CR0_NW) && !(new_val & X86_CR0_CD)))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
cr4 = ctxt->ops->get_cr(ctxt, 4);
|
|
|
|
|
ctxt->ops->get_msr(ctxt, MSR_EFER, &efer);
|
2011-04-04 10:39:28 +00:00
|
|
|
|
|
|
|
|
if ((new_val & X86_CR0_PG) && (efer & EFER_LME) &&
|
|
|
|
|
!(cr4 & X86_CR4_PAE))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case 3: {
|
|
|
|
|
u64 rsvd = 0;
|
|
|
|
|
|
2011-04-20 12:21:35 +00:00
|
|
|
ctxt->ops->get_msr(ctxt, MSR_EFER, &efer);
|
|
|
|
|
if (efer & EFER_LMA)
|
2011-04-04 10:39:28 +00:00
|
|
|
rsvd = CR3_L_MODE_RESERVED_BITS;
|
2011-04-20 12:24:32 +00:00
|
|
|
else if (ctxt->ops->get_cr(ctxt, 4) & X86_CR4_PAE)
|
2011-04-04 10:39:28 +00:00
|
|
|
rsvd = CR3_PAE_RESERVED_BITS;
|
2011-04-20 12:24:32 +00:00
|
|
|
else if (ctxt->ops->get_cr(ctxt, 0) & X86_CR0_PG)
|
2011-04-04 10:39:28 +00:00
|
|
|
rsvd = CR3_NONPAE_RESERVED_BITS;
|
|
|
|
|
|
|
|
|
|
if (new_val & rsvd)
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case 4: {
|
2011-04-20 10:37:53 +00:00
|
|
|
ctxt->ops->get_msr(ctxt, MSR_EFER, &efer);
|
2011-04-04 10:39:28 +00:00
|
|
|
|
|
|
|
|
if ((efer & EFER_LMA) && !(new_val & X86_CR4_PAE))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:29 +00:00
|
|
|
static int check_dr7_gd(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
unsigned long dr7;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ctxt->ops->get_dr(ctxt, 7, &dr7);
|
2011-04-04 10:39:29 +00:00
|
|
|
|
|
|
|
|
/* Check if DR7.Global_Enable is set */
|
|
|
|
|
return dr7 & (1 << 13);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int check_dr_read(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
int dr = ctxt->modrm_reg;
|
2011-04-04 10:39:29 +00:00
|
|
|
u64 cr4;
|
|
|
|
|
|
|
|
|
|
if (dr > 7)
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
cr4 = ctxt->ops->get_cr(ctxt, 4);
|
2011-04-04 10:39:29 +00:00
|
|
|
if ((cr4 & X86_CR4_DE) && (dr == 4 || dr == 5))
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
|
|
|
|
if (check_dr7_gd(ctxt))
|
|
|
|
|
return emulate_db(ctxt);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int check_dr_write(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
u64 new_val = ctxt->src.val64;
|
|
|
|
|
int dr = ctxt->modrm_reg;
|
2011-04-04 10:39:29 +00:00
|
|
|
|
|
|
|
|
if ((dr == 6 || dr == 7) && (new_val & 0xffffffff00000000ULL))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
return check_dr_read(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:31 +00:00
|
|
|
static int check_svme(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
u64 efer;
|
|
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
ctxt->ops->get_msr(ctxt, MSR_EFER, &efer);
|
2011-04-04 10:39:31 +00:00
|
|
|
|
|
|
|
|
if (!(efer & EFER_SVME))
|
|
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int check_svme_pa(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2012-08-27 20:46:17 +00:00
|
|
|
u64 rax = reg_read(ctxt, VCPU_REGS_RAX);
|
2011-04-04 10:39:31 +00:00
|
|
|
|
|
|
|
|
/* Valid physical address? */
|
2011-04-21 16:09:22 +00:00
|
|
|
if (rax & 0xffff000000000000ULL)
|
2011-04-04 10:39:31 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
return check_svme(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:32 +00:00
|
|
|
static int check_rdtsc(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-04-20 10:37:53 +00:00
|
|
|
u64 cr4 = ctxt->ops->get_cr(ctxt, 4);
|
2011-04-04 10:39:32 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
if (cr4 & X86_CR4_TSD && ctxt->ops->cpl(ctxt))
|
2011-04-04 10:39:32 +00:00
|
|
|
return emulate_ud(ctxt);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:33 +00:00
|
|
|
static int check_rdpmc(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-04-20 10:37:53 +00:00
|
|
|
u64 cr4 = ctxt->ops->get_cr(ctxt, 4);
|
2012-08-27 20:46:17 +00:00
|
|
|
u64 rcx = reg_read(ctxt, VCPU_REGS_RCX);
|
2011-04-04 10:39:33 +00:00
|
|
|
|
2011-04-20 10:37:53 +00:00
|
|
|
if ((!(cr4 & X86_CR4_PCE) && ctxt->ops->cpl(ctxt)) ||
|
2011-04-04 10:39:33 +00:00
|
|
|
(rcx > 3))
|
|
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:35 +00:00
|
|
|
static int check_perm_in(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.bytes = min(ctxt->dst.bytes, 4u);
|
|
|
|
|
if (!emulator_io_permited(ctxt, ctxt->src.val, ctxt->dst.bytes))
|
2011-04-04 10:39:35 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int check_perm_out(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.bytes = min(ctxt->src.bytes, 4u);
|
|
|
|
|
if (!emulator_io_permited(ctxt, ctxt->dst.val, ctxt->src.bytes))
|
2011-04-04 10:39:35 +00:00
|
|
|
return emulate_gp(ctxt, 0);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
2010-07-29 12:11:53 +00:00
|
|
|
#define D(_y) { .flags = (_y) }
|
2011-04-04 10:39:22 +00:00
|
|
|
#define DI(_y, _i) { .flags = (_y), .intercept = x86_intercept_##_i }
|
2011-04-04 10:39:25 +00:00
|
|
|
#define DIP(_y, _i, _p) { .flags = (_y), .intercept = x86_intercept_##_i, \
|
|
|
|
|
.check_perm = (_p) }
|
2010-07-29 12:11:53 +00:00
|
|
|
#define N D(0)
|
2011-04-04 10:39:31 +00:00
|
|
|
#define EXT(_f, _e) { .flags = ((_f) | RMExt), .u.group = (_e) }
|
2012-04-30 08:46:31 +00:00
|
|
|
#define G(_f, _g) { .flags = ((_f) | Group | ModRM), .u.group = (_g) }
|
|
|
|
|
#define GD(_f, _g) { .flags = ((_f) | GroupDual | ModRM), .u.gdual = (_g) }
|
2010-07-29 12:11:53 +00:00
|
|
|
#define I(_f, _e) { .flags = (_f), .u.execute = (_e) }
|
2011-04-04 10:39:22 +00:00
|
|
|
#define II(_f, _e, _i) \
|
|
|
|
|
{ .flags = (_f), .u.execute = (_e), .intercept = x86_intercept_##_i }
|
2011-04-04 10:39:25 +00:00
|
|
|
#define IIP(_f, _e, _i, _p) \
|
|
|
|
|
{ .flags = (_f), .u.execute = (_e), .intercept = x86_intercept_##_i, \
|
|
|
|
|
.check_perm = (_p) }
|
2010-01-20 16:09:23 +00:00
|
|
|
#define GP(_f, _g) { .flags = ((_f) | Prefix), .u.gprefix = (_g) }
|
2010-07-29 12:11:53 +00:00
|
|
|
|
2010-08-26 08:56:06 +00:00
|
|
|
#define D2bv(_f) D((_f) | ByteOp), D(_f)
|
2011-04-04 10:39:35 +00:00
|
|
|
#define D2bvIP(_f, _i, _p) DIP((_f) | ByteOp, _i, _p), DIP(_f, _i, _p)
|
2010-08-26 08:56:06 +00:00
|
|
|
#define I2bv(_f, _e) I((_f) | ByteOp, _e), I(_f, _e)
|
2011-11-22 06:16:54 +00:00
|
|
|
#define I2bvIP(_f, _e, _i, _p) \
|
|
|
|
|
IIP((_f) | ByteOp, _e, _i, _p), IIP(_f, _e, _i, _p)
|
2010-08-26 08:56:06 +00:00
|
|
|
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
#define I6ALU(_f, _e) I2bv((_f) | DstMem | SrcReg | ModRM, _e), \
|
|
|
|
|
I2bv(((_f) | DstReg | SrcMem | ModRM) & ~Lock, _e), \
|
|
|
|
|
I2bv(((_f) & ~Lock) | DstAcc | SrcImm, _e)
|
2010-08-26 15:34:55 +00:00
|
|
|
|
2011-04-04 10:39:32 +00:00
|
|
|
static struct opcode group7_rm1[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
DI(SrcNone | Priv, monitor),
|
|
|
|
|
DI(SrcNone | Priv, mwait),
|
2011-04-04 10:39:32 +00:00
|
|
|
N, N, N, N, N, N,
|
|
|
|
|
};
|
|
|
|
|
|
2011-04-04 10:39:31 +00:00
|
|
|
static struct opcode group7_rm3[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
DIP(SrcNone | Prot | Priv, vmrun, check_svme_pa),
|
|
|
|
|
II(SrcNone | Prot | VendorSpecific, em_vmmcall, vmmcall),
|
|
|
|
|
DIP(SrcNone | Prot | Priv, vmload, check_svme_pa),
|
|
|
|
|
DIP(SrcNone | Prot | Priv, vmsave, check_svme_pa),
|
|
|
|
|
DIP(SrcNone | Prot | Priv, stgi, check_svme),
|
|
|
|
|
DIP(SrcNone | Prot | Priv, clgi, check_svme),
|
|
|
|
|
DIP(SrcNone | Prot | Priv, skinit, check_svme),
|
|
|
|
|
DIP(SrcNone | Prot | Priv, invlpga, check_svme),
|
2011-04-04 10:39:31 +00:00
|
|
|
};
|
2010-08-26 15:34:55 +00:00
|
|
|
|
2011-04-04 10:39:32 +00:00
|
|
|
static struct opcode group7_rm7[] = {
|
|
|
|
|
N,
|
2012-04-30 08:46:31 +00:00
|
|
|
DIP(SrcNone, rdtscp, check_rdtsc),
|
2011-04-04 10:39:32 +00:00
|
|
|
N, N, N, N, N, N,
|
|
|
|
|
};
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
|
2010-07-29 12:11:53 +00:00
|
|
|
static struct opcode group1[] = {
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I(Lock, em_add),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I(Lock | PageTable, em_or),
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I(Lock, em_adc),
|
|
|
|
|
I(Lock, em_sbb),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I(Lock | PageTable, em_and),
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I(Lock, em_sub),
|
|
|
|
|
I(Lock, em_xor),
|
|
|
|
|
I(0, em_cmp),
|
2010-07-29 12:11:53 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct opcode group1A[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
I(DstMem | SrcNone | Mov | Stack, em_pop), N, N, N, N, N, N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct opcode group3[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
I(DstMem | SrcImm, em_test),
|
|
|
|
|
I(DstMem | SrcImm, em_test),
|
|
|
|
|
I(DstMem | SrcNone | Lock, em_not),
|
|
|
|
|
I(DstMem | SrcNone | Lock, em_neg),
|
|
|
|
|
I(SrcMem, em_mul_ex),
|
|
|
|
|
I(SrcMem, em_imul_ex),
|
|
|
|
|
I(SrcMem, em_div_ex),
|
|
|
|
|
I(SrcMem, em_idiv_ex),
|
2010-07-29 12:11:53 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct opcode group4[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
I(ByteOp | DstMem | SrcNone | Lock, em_grp45),
|
|
|
|
|
I(ByteOp | DstMem | SrcNone | Lock, em_grp45),
|
2010-07-29 12:11:53 +00:00
|
|
|
N, N, N, N, N, N,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct opcode group5[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
I(DstMem | SrcNone | Lock, em_grp45),
|
|
|
|
|
I(DstMem | SrcNone | Lock, em_grp45),
|
|
|
|
|
I(SrcMem | Stack, em_grp45),
|
|
|
|
|
I(SrcMemFAddr | ImplicitOps | Stack, em_call_far),
|
|
|
|
|
I(SrcMem | Stack, em_grp45),
|
|
|
|
|
I(SrcMemFAddr | ImplicitOps, em_grp45),
|
|
|
|
|
I(SrcMem | Stack, em_grp45), N,
|
2010-07-29 12:11:53 +00:00
|
|
|
};
|
|
|
|
|
|
2011-04-04 10:39:30 +00:00
|
|
|
static struct opcode group6[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
DI(Prot, sldt),
|
|
|
|
|
DI(Prot, str),
|
2012-06-13 09:28:33 +00:00
|
|
|
II(Prot | Priv | SrcMem16, em_lldt, lldt),
|
2012-06-13 13:33:29 +00:00
|
|
|
II(Prot | Priv | SrcMem16, em_ltr, ltr),
|
2011-04-04 10:39:30 +00:00
|
|
|
N, N, N, N,
|
|
|
|
|
};
|
|
|
|
|
|
2010-07-29 12:11:53 +00:00
|
|
|
static struct group_dual group7 = { {
|
2012-06-10 14:21:18 +00:00
|
|
|
II(Mov | DstMem | Priv, em_sgdt, sgdt),
|
|
|
|
|
II(Mov | DstMem | Priv, em_sidt, sidt),
|
2012-04-30 08:46:31 +00:00
|
|
|
II(SrcMem | Priv, em_lgdt, lgdt),
|
|
|
|
|
II(SrcMem | Priv, em_lidt, lidt),
|
|
|
|
|
II(SrcNone | DstMem | Mov, em_smsw, smsw), N,
|
|
|
|
|
II(SrcMem16 | Mov | Priv, em_lmsw, lmsw),
|
|
|
|
|
II(SrcMem | ByteOp | Priv | NoAccess, em_invlpg, invlpg),
|
2010-07-29 12:11:53 +00:00
|
|
|
}, {
|
2012-04-30 08:46:31 +00:00
|
|
|
I(SrcNone | Priv | VendorSpecific, em_vmcall),
|
2011-04-21 09:21:50 +00:00
|
|
|
EXT(0, group7_rm1),
|
2011-04-04 10:39:31 +00:00
|
|
|
N, EXT(0, group7_rm3),
|
2012-04-30 08:46:31 +00:00
|
|
|
II(SrcNone | DstMem | Mov, em_smsw, smsw), N,
|
|
|
|
|
II(SrcMem16 | Mov | Priv, em_lmsw, lmsw),
|
|
|
|
|
EXT(0, group7_rm7),
|
2010-07-29 12:11:53 +00:00
|
|
|
} };
|
|
|
|
|
|
|
|
|
|
static struct opcode group8[] = {
|
|
|
|
|
N, N, N, N,
|
2012-04-30 08:46:31 +00:00
|
|
|
I(DstMem | SrcImmByte, em_bt),
|
|
|
|
|
I(DstMem | SrcImmByte | Lock | PageTable, em_bts),
|
|
|
|
|
I(DstMem | SrcImmByte | Lock, em_btr),
|
|
|
|
|
I(DstMem | SrcImmByte | Lock | PageTable, em_btc),
|
2010-07-29 12:11:53 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct group_dual group9 = { {
|
2012-04-30 08:46:31 +00:00
|
|
|
N, I(DstMem64 | Lock | PageTable, em_cmpxchg8b), N, N, N, N, N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
}, {
|
|
|
|
|
N, N, N, N, N, N, N, N,
|
|
|
|
|
} };
|
|
|
|
|
|
2010-08-03 12:05:46 +00:00
|
|
|
static struct opcode group11[] = {
|
2012-04-30 08:46:31 +00:00
|
|
|
I(DstMem | SrcImm | Mov | PageTable, em_mov),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
X7(D(Undefined)),
|
2010-08-03 12:05:46 +00:00
|
|
|
};
|
|
|
|
|
|
2010-01-20 16:09:23 +00:00
|
|
|
static struct gprefix pfx_0f_6f_0f_7f = {
|
2012-04-09 15:40:03 +00:00
|
|
|
I(Mmx, em_mov), I(Sse | Aligned, em_mov), N, I(Sse | Unaligned, em_mov),
|
2010-01-20 16:09:23 +00:00
|
|
|
};
|
|
|
|
|
|
2012-04-09 15:40:01 +00:00
|
|
|
static struct gprefix pfx_vmovntpx = {
|
|
|
|
|
I(0, em_mov), N, N, N,
|
|
|
|
|
};
|
|
|
|
|
|
2010-07-29 12:11:53 +00:00
|
|
|
static struct opcode opcode_table[256] = {
|
|
|
|
|
/* 0x00 - 0x07 */
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I6ALU(Lock, em_add),
|
2011-09-13 07:45:51 +00:00
|
|
|
I(ImplicitOps | Stack | No64 | Src2ES, em_push_sreg),
|
|
|
|
|
I(ImplicitOps | Stack | No64 | Src2ES, em_pop_sreg),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x08 - 0x0F */
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I6ALU(Lock | PageTable, em_or),
|
2011-09-13 07:45:51 +00:00
|
|
|
I(ImplicitOps | Stack | No64 | Src2CS, em_push_sreg),
|
|
|
|
|
N,
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x10 - 0x17 */
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I6ALU(Lock, em_adc),
|
2011-09-13 07:45:51 +00:00
|
|
|
I(ImplicitOps | Stack | No64 | Src2SS, em_push_sreg),
|
|
|
|
|
I(ImplicitOps | Stack | No64 | Src2SS, em_pop_sreg),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x18 - 0x1F */
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I6ALU(Lock, em_sbb),
|
2011-09-13 07:45:51 +00:00
|
|
|
I(ImplicitOps | Stack | No64 | Src2DS, em_push_sreg),
|
|
|
|
|
I(ImplicitOps | Stack | No64 | Src2DS, em_pop_sreg),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x20 - 0x27 */
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I6ALU(Lock | PageTable, em_and), N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x28 - 0x2F */
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I6ALU(Lock, em_sub), N, I(ByteOp | DstAcc | No64, em_das),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x30 - 0x37 */
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I6ALU(Lock, em_xor), N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x38 - 0x3F */
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I6ALU(0, em_cmp), N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x40 - 0x4F */
|
|
|
|
|
X16(D(DstReg)),
|
|
|
|
|
/* 0x50 - 0x57 */
|
2010-07-29 12:11:55 +00:00
|
|
|
X8(I(SrcReg | Stack, em_push)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x58 - 0x5F */
|
2011-04-23 09:49:40 +00:00
|
|
|
X8(I(DstReg | Stack, em_pop)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x60 - 0x67 */
|
2011-04-23 09:51:07 +00:00
|
|
|
I(ImplicitOps | Stack | No64, em_pusha),
|
|
|
|
|
I(ImplicitOps | Stack | No64, em_popa),
|
2010-07-29 12:11:53 +00:00
|
|
|
N, D(DstReg | SrcMem32 | ModRM | Mov) /* movsxd (x86/64) */ ,
|
|
|
|
|
N, N, N, N,
|
|
|
|
|
/* 0x68 - 0x6F */
|
2010-08-18 16:29:33 +00:00
|
|
|
I(SrcImm | Mov | Stack, em_push),
|
|
|
|
|
I(DstReg | SrcMem | ModRM | Src2Imm, em_imul_3op),
|
2010-08-18 15:25:25 +00:00
|
|
|
I(SrcImmByte | Mov | Stack, em_push),
|
|
|
|
|
I(DstReg | SrcMem | ModRM | Src2ImmByte, em_imul_3op),
|
2011-11-23 03:27:39 +00:00
|
|
|
I2bvIP(DstDI | SrcDX | Mov | String, em_in, ins, check_perm_in), /* insb, insw/insd */
|
|
|
|
|
I2bvIP(SrcSI | DstDX | String, em_out, outs, check_perm_out), /* outsb, outsw/outsd */
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x70 - 0x7F */
|
|
|
|
|
X16(D(SrcImmByte)),
|
|
|
|
|
/* 0x80 - 0x87 */
|
2012-04-30 08:46:31 +00:00
|
|
|
G(ByteOp | DstMem | SrcImm, group1),
|
|
|
|
|
G(DstMem | SrcImm, group1),
|
|
|
|
|
G(ByteOp | DstMem | SrcImm | No64, group1),
|
|
|
|
|
G(DstMem | SrcImmByte, group1),
|
2011-05-29 12:57:53 +00:00
|
|
|
I2bv(DstMem | SrcReg | ModRM, em_test),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I2bv(DstMem | SrcReg | ModRM | Lock | PageTable, em_xchg),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x88 - 0x8F */
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I2bv(DstMem | SrcReg | ModRM | Mov | PageTable, em_mov),
|
2010-08-03 11:46:56 +00:00
|
|
|
I2bv(DstReg | SrcMem | ModRM | Mov, em_mov),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I(DstMem | SrcNone | ModRM | Mov | PageTable, em_mov_rm_sreg),
|
2011-05-29 13:01:33 +00:00
|
|
|
D(ModRM | SrcMem | NoAccess | DstReg),
|
|
|
|
|
I(ImplicitOps | SrcMem16 | ModRM, em_mov_sreg_rm),
|
|
|
|
|
G(0, group1A),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x90 - 0x97 */
|
2011-04-04 10:39:34 +00:00
|
|
|
DI(SrcAcc | DstReg, pause), X7(D(SrcAcc | DstReg)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x98 - 0x9F */
|
2010-08-19 12:13:00 +00:00
|
|
|
D(DstAcc | SrcNone), I(ImplicitOps | SrcAcc, em_cwd),
|
2010-08-25 06:10:53 +00:00
|
|
|
I(SrcImmFAddr | No64, em_call_far), N,
|
2011-04-23 09:52:56 +00:00
|
|
|
II(ImplicitOps | Stack, em_pushf, pushf),
|
2012-06-11 10:09:07 +00:00
|
|
|
II(ImplicitOps | Stack, em_popf, popf), N, I(ImplicitOps, em_lahf),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xA0 - 0xA7 */
|
2010-08-03 11:46:56 +00:00
|
|
|
I2bv(DstAcc | SrcMem | Mov | MemAbs, em_mov),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
I2bv(DstMem | SrcAcc | Mov | MemAbs | PageTable, em_mov),
|
2010-08-03 11:46:56 +00:00
|
|
|
I2bv(SrcSI | DstDI | Mov | String, em_mov),
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I2bv(SrcSI | DstDI | String, em_cmp),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xA8 - 0xAF */
|
2011-05-29 12:57:53 +00:00
|
|
|
I2bv(DstAcc | SrcImm, em_test),
|
2010-08-03 11:46:56 +00:00
|
|
|
I2bv(SrcAcc | DstDI | Mov | String, em_mov),
|
|
|
|
|
I2bv(SrcSI | DstAcc | Mov | String, em_mov),
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
I2bv(SrcAcc | DstDI | String, em_cmp),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xB0 - 0xB7 */
|
2010-08-03 11:46:56 +00:00
|
|
|
X8(I(ByteOp | DstReg | SrcImm | Mov, em_mov)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xB8 - 0xBF */
|
2010-08-03 11:46:56 +00:00
|
|
|
X8(I(DstReg | SrcImm | Mov, em_mov)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xC0 - 0xC7 */
|
2010-08-26 08:56:11 +00:00
|
|
|
D2bv(DstMem | SrcImmByte | ModRM),
|
2010-08-18 12:12:09 +00:00
|
|
|
I(ImplicitOps | Stack | SrcImmU16, em_ret_near_imm),
|
2011-05-29 13:00:22 +00:00
|
|
|
I(ImplicitOps | Stack, em_ret),
|
2011-09-13 07:45:50 +00:00
|
|
|
I(DstReg | SrcMemFAddr | ModRM | No64 | Src2ES, em_lseg),
|
|
|
|
|
I(DstReg | SrcMemFAddr | ModRM | No64 | Src2DS, em_lseg),
|
2010-08-03 12:05:46 +00:00
|
|
|
G(ByteOp, group11), G(0, group11),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xC8 - 0xCF */
|
2012-06-12 17:03:23 +00:00
|
|
|
I(Stack | SrcImmU16 | Src2ImmByte, em_enter), I(Stack, em_leave),
|
|
|
|
|
N, I(ImplicitOps | Stack, em_ret_far),
|
2011-04-04 10:39:23 +00:00
|
|
|
D(ImplicitOps), DI(SrcImmByte, intn),
|
2011-05-29 12:56:26 +00:00
|
|
|
D(ImplicitOps | No64), II(ImplicitOps, em_iret, iret),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xD0 - 0xD7 */
|
2010-08-26 08:56:11 +00:00
|
|
|
D2bv(DstMem | SrcOne | ModRM), D2bv(DstMem | ModRM),
|
2010-07-29 12:11:53 +00:00
|
|
|
N, N, N, N,
|
|
|
|
|
/* 0xD8 - 0xDF */
|
|
|
|
|
N, N, N, N, N, N, N, N,
|
|
|
|
|
/* 0xE0 - 0xE7 */
|
2011-05-29 13:04:08 +00:00
|
|
|
X3(I(SrcImmByte, em_loop)),
|
|
|
|
|
I(SrcImmByte, em_jcxz),
|
2011-11-22 06:16:54 +00:00
|
|
|
I2bvIP(SrcImmUByte | DstAcc, em_in, in, check_perm_in),
|
|
|
|
|
I2bvIP(SrcAcc | DstImmUByte, em_out, out, check_perm_out),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xE8 - 0xEF */
|
2011-11-22 06:18:35 +00:00
|
|
|
I(SrcImm | Stack, em_call), D(SrcImm | ImplicitOps),
|
2011-05-29 12:56:26 +00:00
|
|
|
I(SrcImmFAddr | No64, em_jmp_far), D(SrcImmByte | ImplicitOps),
|
2011-11-22 06:16:54 +00:00
|
|
|
I2bvIP(SrcDX | DstAcc, em_in, in, check_perm_in),
|
|
|
|
|
I2bvIP(SrcAcc | DstDX, em_out, out, check_perm_out),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xF0 - 0xF7 */
|
2011-04-04 10:39:34 +00:00
|
|
|
N, DI(ImplicitOps, icebp), N, N,
|
2011-04-04 10:39:23 +00:00
|
|
|
DI(ImplicitOps | Priv, hlt), D(ImplicitOps),
|
|
|
|
|
G(ByteOp, group3), G(0, group3),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xF8 - 0xFF */
|
2011-05-29 13:05:15 +00:00
|
|
|
D(ImplicitOps), D(ImplicitOps),
|
|
|
|
|
I(ImplicitOps, em_cli), I(ImplicitOps, em_sti),
|
2010-07-29 12:11:53 +00:00
|
|
|
D(ImplicitOps), D(ImplicitOps), G(0, group4), G(0, group5),
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct opcode twobyte_table[256] = {
|
|
|
|
|
/* 0x00 - 0x0F */
|
2011-04-04 10:39:30 +00:00
|
|
|
G(0, group6), GD(0, &group7), N, N,
|
2011-05-29 12:56:26 +00:00
|
|
|
N, I(ImplicitOps | VendorSpecific, em_syscall),
|
|
|
|
|
II(ImplicitOps | Priv, em_clts, clts), N,
|
2011-04-04 10:39:23 +00:00
|
|
|
DI(ImplicitOps | Priv, invd), DI(ImplicitOps | Priv, wbinvd), N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
N, D(ImplicitOps | ModRM), N, N,
|
|
|
|
|
/* 0x10 - 0x1F */
|
|
|
|
|
N, N, N, N, N, N, N, N, D(ImplicitOps | ModRM), N, N, N, N, N, N, N,
|
|
|
|
|
/* 0x20 - 0x2F */
|
2011-04-04 10:39:28 +00:00
|
|
|
DIP(ModRM | DstMem | Priv | Op3264, cr_read, check_cr_read),
|
2011-04-04 10:39:29 +00:00
|
|
|
DIP(ModRM | DstMem | Priv | Op3264, dr_read, check_dr_read),
|
2011-11-22 06:19:19 +00:00
|
|
|
IIP(ModRM | SrcMem | Priv | Op3264, em_cr_write, cr_write, check_cr_write),
|
|
|
|
|
IIP(ModRM | SrcMem | Priv | Op3264, em_dr_write, dr_write, check_dr_write),
|
2010-07-29 12:11:53 +00:00
|
|
|
N, N, N, N,
|
2012-04-09 15:40:01 +00:00
|
|
|
N, N, N, GP(ModRM | DstMem | SrcReg | Sse | Mov | Aligned, &pfx_vmovntpx),
|
|
|
|
|
N, N, N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x30 - 0x3F */
|
2011-11-22 06:20:03 +00:00
|
|
|
II(ImplicitOps | Priv, em_wrmsr, wrmsr),
|
2011-04-04 10:39:33 +00:00
|
|
|
IIP(ImplicitOps, em_rdtsc, rdtsc, check_rdtsc),
|
2011-11-22 06:20:03 +00:00
|
|
|
II(ImplicitOps | Priv, em_rdmsr, rdmsr),
|
2011-11-10 12:57:30 +00:00
|
|
|
IIP(ImplicitOps, em_rdpmc, rdpmc, check_rdpmc),
|
2011-05-29 12:56:26 +00:00
|
|
|
I(ImplicitOps | VendorSpecific, em_sysenter),
|
|
|
|
|
I(ImplicitOps | Priv | VendorSpecific, em_sysexit),
|
2011-02-01 14:32:03 +00:00
|
|
|
N, N,
|
2010-07-29 12:11:53 +00:00
|
|
|
N, N, N, N, N, N, N, N,
|
|
|
|
|
/* 0x40 - 0x4F */
|
|
|
|
|
X16(D(DstReg | SrcMem | ModRM | Mov)),
|
|
|
|
|
/* 0x50 - 0x5F */
|
|
|
|
|
N, N, N, N, N, N, N, N, N, N, N, N, N, N, N, N,
|
|
|
|
|
/* 0x60 - 0x6F */
|
2010-01-20 16:09:23 +00:00
|
|
|
N, N, N, N,
|
|
|
|
|
N, N, N, N,
|
|
|
|
|
N, N, N, N,
|
|
|
|
|
N, N, N, GP(SrcMem | DstReg | ModRM | Mov, &pfx_0f_6f_0f_7f),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x70 - 0x7F */
|
2010-01-20 16:09:23 +00:00
|
|
|
N, N, N, N,
|
|
|
|
|
N, N, N, N,
|
|
|
|
|
N, N, N, N,
|
|
|
|
|
N, N, N, GP(SrcReg | DstMem | ModRM | Mov, &pfx_0f_6f_0f_7f),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0x80 - 0x8F */
|
|
|
|
|
X16(D(SrcImm)),
|
|
|
|
|
/* 0x90 - 0x9F */
|
2010-08-06 09:10:07 +00:00
|
|
|
X16(D(ByteOp | DstMem | SrcNone | ModRM| Mov)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xA0 - 0xA7 */
|
2011-09-13 07:45:51 +00:00
|
|
|
I(Stack | Src2FS, em_push_sreg), I(Stack | Src2FS, em_pop_sreg),
|
2012-06-07 11:11:36 +00:00
|
|
|
II(ImplicitOps, em_cpuid, cpuid), I(DstMem | SrcReg | ModRM | BitOp, em_bt),
|
2010-07-29 12:11:53 +00:00
|
|
|
D(DstMem | SrcReg | Src2ImmByte | ModRM),
|
|
|
|
|
D(DstMem | SrcReg | Src2CL | ModRM), N, N,
|
|
|
|
|
/* 0xA8 - 0xAF */
|
2011-09-13 07:45:51 +00:00
|
|
|
I(Stack | Src2GS, em_push_sreg), I(Stack | Src2GS, em_pop_sreg),
|
KVM: x86: tag the instructions which are used to write page table
The idea is from Avi:
| tag instructions that are typically used to modify the page tables, and
| drop shadow if any other instruction is used.
| The list would include, I'd guess, and, or, bts, btc, mov, xchg, cmpxchg,
| and cmpxchg8b.
This patch is used to tag the instructions and in the later path, shadow page
is dropped if it is written by other instructions
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-09-22 08:53:46 +00:00
|
|
|
DI(ImplicitOps, rsm),
|
2011-11-22 06:17:48 +00:00
|
|
|
I(DstMem | SrcReg | ModRM | BitOp | Lock | PageTable, em_bts),
|
2010-07-29 12:11:53 +00:00
|
|
|
D(DstMem | SrcReg | Src2ImmByte | ModRM),
|
|
|
|
|
D(DstMem | SrcReg | Src2CL | ModRM),
|
2010-08-18 15:31:43 +00:00
|
|
|
D(ModRM), I(DstReg | SrcMem | ModRM, em_imul),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xB0 - 0xB7 */
|
2011-11-22 06:20:47 +00:00
|
|
|
I2bv(DstMem | SrcReg | ModRM | Lock | PageTable, em_cmpxchg),
|
2011-09-13 07:45:50 +00:00
|
|
|
I(DstReg | SrcMemFAddr | ModRM | Src2SS, em_lseg),
|
2011-11-22 06:17:48 +00:00
|
|
|
I(DstMem | SrcReg | ModRM | BitOp | Lock, em_btr),
|
2011-09-13 07:45:50 +00:00
|
|
|
I(DstReg | SrcMemFAddr | ModRM | Src2FS, em_lseg),
|
|
|
|
|
I(DstReg | SrcMemFAddr | ModRM | Src2GS, em_lseg),
|
2012-01-16 13:08:45 +00:00
|
|
|
D(DstReg | SrcMem8 | ModRM | Mov), D(DstReg | SrcMem16 | ModRM | Mov),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xB8 - 0xBF */
|
|
|
|
|
N, N,
|
2011-11-22 06:17:48 +00:00
|
|
|
G(BitOp, group8),
|
|
|
|
|
I(DstMem | SrcReg | ModRM | BitOp | Lock | PageTable, em_btc),
|
2011-11-22 06:21:33 +00:00
|
|
|
I(DstReg | SrcMem | ModRM, em_bsf), I(DstReg | SrcMem | ModRM, em_bsr),
|
2012-01-16 13:08:45 +00:00
|
|
|
D(DstReg | SrcMem8 | ModRM | Mov), D(DstReg | SrcMem16 | ModRM | Mov),
|
2012-06-13 09:25:06 +00:00
|
|
|
/* 0xC0 - 0xC7 */
|
2010-08-26 08:56:13 +00:00
|
|
|
D2bv(DstMem | SrcReg | ModRM | Lock),
|
2010-08-17 01:19:34 +00:00
|
|
|
N, D(DstMem | SrcReg | ModRM | Mov),
|
2010-07-29 12:11:53 +00:00
|
|
|
N, N, N, GD(0, &group9),
|
2012-06-13 09:25:06 +00:00
|
|
|
/* 0xC8 - 0xCF */
|
|
|
|
|
X8(I(DstReg, em_bswap)),
|
2010-07-29 12:11:53 +00:00
|
|
|
/* 0xD0 - 0xDF */
|
|
|
|
|
N, N, N, N, N, N, N, N, N, N, N, N, N, N, N, N,
|
|
|
|
|
/* 0xE0 - 0xEF */
|
|
|
|
|
N, N, N, N, N, N, N, N, N, N, N, N, N, N, N, N,
|
|
|
|
|
/* 0xF0 - 0xFF */
|
|
|
|
|
N, N, N, N, N, N, N, N, N, N, N, N, N, N, N, N
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
#undef D
|
|
|
|
|
#undef N
|
|
|
|
|
#undef G
|
|
|
|
|
#undef GD
|
|
|
|
|
#undef I
|
2010-01-20 16:09:23 +00:00
|
|
|
#undef GP
|
2011-04-04 10:39:31 +00:00
|
|
|
#undef EXT
|
2010-07-29 12:11:53 +00:00
|
|
|
|
2010-08-26 08:56:06 +00:00
|
|
|
#undef D2bv
|
2011-04-04 10:39:35 +00:00
|
|
|
#undef D2bvIP
|
2010-08-26 08:56:06 +00:00
|
|
|
#undef I2bv
|
2011-11-22 06:16:54 +00:00
|
|
|
#undef I2bvIP
|
KVM: x86 emulator: Use opcode::execute for Group 1, CMPS and SCAS
The following instructions are changed to use opcode::execute.
Group 1 (80-83)
ADD (00-05), OR (08-0D), ADC (10-15), SBB (18-1D), AND (20-25),
SUB (28-2D), XOR (30-35), CMP (38-3D)
CMPS (A6-A7), SCAS (AE-AF)
The last two do the same as CMP in the emulator, so em_cmp() is used.
Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Avi Kivity <avi@redhat.com>
2011-04-23 09:48:02 +00:00
|
|
|
#undef I6ALU
|
2010-08-26 08:56:06 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
static unsigned imm_size(struct x86_emulate_ctxt *ctxt)
|
2010-08-18 16:20:21 +00:00
|
|
|
{
|
|
|
|
|
unsigned size;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
size = (ctxt->d & ByteOp) ? 1 : ctxt->op_bytes;
|
2010-08-18 16:20:21 +00:00
|
|
|
if (size == 8)
|
|
|
|
|
size = 4;
|
|
|
|
|
return size;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int decode_imm(struct x86_emulate_ctxt *ctxt, struct operand *op,
|
|
|
|
|
unsigned size, bool sign_extension)
|
|
|
|
|
{
|
|
|
|
|
int rc = X86EMUL_CONTINUE;
|
|
|
|
|
|
|
|
|
|
op->type = OP_IMM;
|
|
|
|
|
op->bytes = size;
|
2011-06-01 12:34:25 +00:00
|
|
|
op->addr.mem.ea = ctxt->_eip;
|
2010-08-18 16:20:21 +00:00
|
|
|
/* NB. Immediates are sign-extended as necessary. */
|
|
|
|
|
switch (op->bytes) {
|
|
|
|
|
case 1:
|
2011-07-30 09:01:26 +00:00
|
|
|
op->val = insn_fetch(s8, ctxt);
|
2010-08-18 16:20:21 +00:00
|
|
|
break;
|
|
|
|
|
case 2:
|
2011-07-30 09:01:26 +00:00
|
|
|
op->val = insn_fetch(s16, ctxt);
|
2010-08-18 16:20:21 +00:00
|
|
|
break;
|
|
|
|
|
case 4:
|
2011-07-30 09:01:26 +00:00
|
|
|
op->val = insn_fetch(s32, ctxt);
|
2010-08-18 16:20:21 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
if (!sign_extension) {
|
|
|
|
|
switch (op->bytes) {
|
|
|
|
|
case 1:
|
|
|
|
|
op->val &= 0xff;
|
|
|
|
|
break;
|
|
|
|
|
case 2:
|
|
|
|
|
op->val &= 0xffff;
|
|
|
|
|
break;
|
|
|
|
|
case 4:
|
|
|
|
|
op->val &= 0xffffffff;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
done:
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2011-09-13 07:45:41 +00:00
|
|
|
static int decode_operand(struct x86_emulate_ctxt *ctxt, struct operand *op,
|
|
|
|
|
unsigned d)
|
|
|
|
|
{
|
|
|
|
|
int rc = X86EMUL_CONTINUE;
|
|
|
|
|
|
|
|
|
|
switch (d) {
|
|
|
|
|
case OpReg:
|
2012-01-16 13:08:45 +00:00
|
|
|
decode_register_operand(ctxt, op);
|
2011-09-13 07:45:41 +00:00
|
|
|
break;
|
|
|
|
|
case OpImmUByte:
|
2011-09-13 07:45:45 +00:00
|
|
|
rc = decode_imm(ctxt, op, 1, false);
|
2011-09-13 07:45:41 +00:00
|
|
|
break;
|
|
|
|
|
case OpMem:
|
2011-09-13 07:45:48 +00:00
|
|
|
ctxt->memop.bytes = (ctxt->d & ByteOp) ? 1 : ctxt->op_bytes;
|
2011-09-13 07:45:47 +00:00
|
|
|
mem_common:
|
|
|
|
|
*op = ctxt->memop;
|
|
|
|
|
ctxt->memopp = op;
|
|
|
|
|
if ((ctxt->d & BitOp) && op == &ctxt->dst)
|
2011-09-13 07:45:41 +00:00
|
|
|
fetch_bit_operand(ctxt);
|
|
|
|
|
op->orig_val = op->val;
|
|
|
|
|
break;
|
2011-09-13 07:45:48 +00:00
|
|
|
case OpMem64:
|
|
|
|
|
ctxt->memop.bytes = 8;
|
|
|
|
|
goto mem_common;
|
2011-09-13 07:45:41 +00:00
|
|
|
case OpAcc:
|
|
|
|
|
op->type = OP_REG;
|
|
|
|
|
op->bytes = (ctxt->d & ByteOp) ? 1 : ctxt->op_bytes;
|
2012-08-27 20:46:17 +00:00
|
|
|
op->addr.reg = reg_rmw(ctxt, VCPU_REGS_RAX);
|
2011-09-13 07:45:41 +00:00
|
|
|
fetch_register_operand(op);
|
|
|
|
|
op->orig_val = op->val;
|
|
|
|
|
break;
|
|
|
|
|
case OpDI:
|
|
|
|
|
op->type = OP_MEM;
|
|
|
|
|
op->bytes = (ctxt->d & ByteOp) ? 1 : ctxt->op_bytes;
|
|
|
|
|
op->addr.mem.ea =
|
2012-08-27 20:46:17 +00:00
|
|
|
register_address(ctxt, reg_read(ctxt, VCPU_REGS_RDI));
|
2011-09-13 07:45:41 +00:00
|
|
|
op->addr.mem.seg = VCPU_SREG_ES;
|
|
|
|
|
op->val = 0;
|
|
|
|
|
break;
|
|
|
|
|
case OpDX:
|
|
|
|
|
op->type = OP_REG;
|
|
|
|
|
op->bytes = 2;
|
2012-08-27 20:46:17 +00:00
|
|
|
op->addr.reg = reg_rmw(ctxt, VCPU_REGS_RDX);
|
2011-09-13 07:45:41 +00:00
|
|
|
fetch_register_operand(op);
|
|
|
|
|
break;
|
2011-09-13 07:45:43 +00:00
|
|
|
case OpCL:
|
|
|
|
|
op->bytes = 1;
|
2012-08-27 20:46:17 +00:00
|
|
|
op->val = reg_read(ctxt, VCPU_REGS_RCX) & 0xff;
|
2011-09-13 07:45:43 +00:00
|
|
|
break;
|
|
|
|
|
case OpImmByte:
|
|
|
|
|
rc = decode_imm(ctxt, op, 1, true);
|
|
|
|
|
break;
|
|
|
|
|
case OpOne:
|
|
|
|
|
op->bytes = 1;
|
|
|
|
|
op->val = 1;
|
|
|
|
|
break;
|
|
|
|
|
case OpImm:
|
|
|
|
|
rc = decode_imm(ctxt, op, imm_size(ctxt), true);
|
|
|
|
|
break;
|
2012-01-16 13:08:44 +00:00
|
|
|
case OpMem8:
|
|
|
|
|
ctxt->memop.bytes = 1;
|
|
|
|
|
goto mem_common;
|
2011-09-13 07:45:47 +00:00
|
|
|
case OpMem16:
|
|
|
|
|
ctxt->memop.bytes = 2;
|
|
|
|
|
goto mem_common;
|
|
|
|
|
case OpMem32:
|
|
|
|
|
ctxt->memop.bytes = 4;
|
|
|
|
|
goto mem_common;
|
|
|
|
|
case OpImmU16:
|
|
|
|
|
rc = decode_imm(ctxt, op, 2, false);
|
|
|
|
|
break;
|
|
|
|
|
case OpImmU:
|
|
|
|
|
rc = decode_imm(ctxt, op, imm_size(ctxt), false);
|
|
|
|
|
break;
|
|
|
|
|
case OpSI:
|
|
|
|
|
op->type = OP_MEM;
|
|
|
|
|
op->bytes = (ctxt->d & ByteOp) ? 1 : ctxt->op_bytes;
|
|
|
|
|
op->addr.mem.ea =
|
2012-08-27 20:46:17 +00:00
|
|
|
register_address(ctxt, reg_read(ctxt, VCPU_REGS_RSI));
|
2011-09-13 07:45:47 +00:00
|
|
|
op->addr.mem.seg = seg_override(ctxt);
|
|
|
|
|
op->val = 0;
|
|
|
|
|
break;
|
|
|
|
|
case OpImmFAddr:
|
|
|
|
|
op->type = OP_IMM;
|
|
|
|
|
op->addr.mem.ea = ctxt->_eip;
|
|
|
|
|
op->bytes = ctxt->op_bytes + 2;
|
|
|
|
|
insn_fetch_arr(op->valptr, op->bytes, ctxt);
|
|
|
|
|
break;
|
|
|
|
|
case OpMemFAddr:
|
|
|
|
|
ctxt->memop.bytes = ctxt->op_bytes + 2;
|
|
|
|
|
goto mem_common;
|
2011-09-13 07:45:49 +00:00
|
|
|
case OpES:
|
|
|
|
|
op->val = VCPU_SREG_ES;
|
|
|
|
|
break;
|
|
|
|
|
case OpCS:
|
|
|
|
|
op->val = VCPU_SREG_CS;
|
|
|
|
|
break;
|
|
|
|
|
case OpSS:
|
|
|
|
|
op->val = VCPU_SREG_SS;
|
|
|
|
|
break;
|
|
|
|
|
case OpDS:
|
|
|
|
|
op->val = VCPU_SREG_DS;
|
|
|
|
|
break;
|
|
|
|
|
case OpFS:
|
|
|
|
|
op->val = VCPU_SREG_FS;
|
|
|
|
|
break;
|
|
|
|
|
case OpGS:
|
|
|
|
|
op->val = VCPU_SREG_GS;
|
|
|
|
|
break;
|
2011-09-13 07:45:41 +00:00
|
|
|
case OpImplicit:
|
|
|
|
|
/* Special instructions do their own operand decoding. */
|
|
|
|
|
default:
|
|
|
|
|
op->type = OP_NONE; /* Disable writeback. */
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
done:
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-14 15:57:43 +00:00
|
|
|
int x86_decode_insn(struct x86_emulate_ctxt *ctxt, void *insn, int insn_len)
|
2010-07-29 12:11:52 +00:00
|
|
|
{
|
|
|
|
|
int rc = X86EMUL_CONTINUE;
|
|
|
|
|
int mode = ctxt->mode;
|
2011-04-24 11:09:59 +00:00
|
|
|
int def_op_bytes, def_ad_bytes, goffset, simd_prefix;
|
2011-03-29 09:34:38 +00:00
|
|
|
bool op_prefix = false;
|
2011-04-24 11:09:59 +00:00
|
|
|
struct opcode opcode;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
2011-09-13 07:45:40 +00:00
|
|
|
ctxt->memop.type = OP_NONE;
|
|
|
|
|
ctxt->memopp = NULL;
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->_eip = ctxt->eip;
|
|
|
|
|
ctxt->fetch.start = ctxt->_eip;
|
|
|
|
|
ctxt->fetch.end = ctxt->fetch.start + insn_len;
|
2010-12-21 10:12:07 +00:00
|
|
|
if (insn_len > 0)
|
2011-06-01 12:34:25 +00:00
|
|
|
memcpy(ctxt->fetch.data, insn, insn_len);
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
switch (mode) {
|
|
|
|
|
case X86EMUL_MODE_REAL:
|
|
|
|
|
case X86EMUL_MODE_VM86:
|
|
|
|
|
case X86EMUL_MODE_PROT16:
|
|
|
|
|
def_op_bytes = def_ad_bytes = 2;
|
|
|
|
|
break;
|
|
|
|
|
case X86EMUL_MODE_PROT32:
|
|
|
|
|
def_op_bytes = def_ad_bytes = 4;
|
|
|
|
|
break;
|
|
|
|
|
#ifdef CONFIG_X86_64
|
|
|
|
|
case X86EMUL_MODE_PROT64:
|
|
|
|
|
def_op_bytes = 4;
|
|
|
|
|
def_ad_bytes = 8;
|
|
|
|
|
break;
|
|
|
|
|
#endif
|
|
|
|
|
default:
|
2011-07-30 09:03:34 +00:00
|
|
|
return EMULATION_FAILED;
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes = def_op_bytes;
|
|
|
|
|
ctxt->ad_bytes = def_ad_bytes;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
/* Legacy prefixes. */
|
|
|
|
|
for (;;) {
|
2011-07-30 09:01:26 +00:00
|
|
|
switch (ctxt->b = insn_fetch(u8, ctxt)) {
|
2010-07-29 12:11:52 +00:00
|
|
|
case 0x66: /* operand-size override */
|
2011-03-29 09:34:38 +00:00
|
|
|
op_prefix = true;
|
2010-07-29 12:11:52 +00:00
|
|
|
/* switch between 2/4 bytes */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes = def_op_bytes ^ 6;
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
case 0x67: /* address-size override */
|
|
|
|
|
if (mode == X86EMUL_MODE_PROT64)
|
|
|
|
|
/* switch between 4/8 bytes */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->ad_bytes = def_ad_bytes ^ 12;
|
2010-07-29 12:11:52 +00:00
|
|
|
else
|
|
|
|
|
/* switch between 2/4 bytes */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->ad_bytes = def_ad_bytes ^ 6;
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
case 0x26: /* ES override */
|
|
|
|
|
case 0x2e: /* CS override */
|
|
|
|
|
case 0x36: /* SS override */
|
|
|
|
|
case 0x3e: /* DS override */
|
2011-06-01 12:34:25 +00:00
|
|
|
set_seg_override(ctxt, (ctxt->b >> 3) & 3);
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
case 0x64: /* FS override */
|
|
|
|
|
case 0x65: /* GS override */
|
2011-06-01 12:34:25 +00:00
|
|
|
set_seg_override(ctxt, ctxt->b & 7);
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
case 0x40 ... 0x4f: /* REX */
|
|
|
|
|
if (mode != X86EMUL_MODE_PROT64)
|
|
|
|
|
goto done_prefixes;
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->rex_prefix = ctxt->b;
|
2010-07-29 12:11:52 +00:00
|
|
|
continue;
|
|
|
|
|
case 0xf0: /* LOCK */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->lock_prefix = 1;
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
case 0xf2: /* REPNE/REPNZ */
|
|
|
|
|
case 0xf3: /* REP/REPE/REPZ */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->rep_prefix = ctxt->b;
|
2010-07-29 12:11:52 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
goto done_prefixes;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Any legacy prefix after a REX prefix nullifies its effect. */
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->rex_prefix = 0;
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
done_prefixes:
|
|
|
|
|
|
|
|
|
|
/* REX prefix. */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->rex_prefix & 8)
|
|
|
|
|
ctxt->op_bytes = 8; /* REX.W */
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
/* Opcode byte(s). */
|
2011-06-01 12:34:25 +00:00
|
|
|
opcode = opcode_table[ctxt->b];
|
2010-08-05 08:34:39 +00:00
|
|
|
/* Two-byte opcode? */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->b == 0x0f) {
|
|
|
|
|
ctxt->twobyte = 1;
|
2011-07-30 09:01:26 +00:00
|
|
|
ctxt->b = insn_fetch(u8, ctxt);
|
2011-06-01 12:34:25 +00:00
|
|
|
opcode = twobyte_table[ctxt->b];
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->d = opcode.flags;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
2012-04-30 08:48:25 +00:00
|
|
|
if (ctxt->d & ModRM)
|
|
|
|
|
ctxt->modrm = insn_fetch(u8, ctxt);
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
while (ctxt->d & GroupMask) {
|
|
|
|
|
switch (ctxt->d & GroupMask) {
|
2011-04-24 11:09:59 +00:00
|
|
|
case Group:
|
2011-06-01 12:34:25 +00:00
|
|
|
goffset = (ctxt->modrm >> 3) & 7;
|
2011-04-24 11:09:59 +00:00
|
|
|
opcode = opcode.u.group[goffset];
|
|
|
|
|
break;
|
|
|
|
|
case GroupDual:
|
2011-06-01 12:34:25 +00:00
|
|
|
goffset = (ctxt->modrm >> 3) & 7;
|
|
|
|
|
if ((ctxt->modrm >> 6) == 3)
|
2011-04-24 11:09:59 +00:00
|
|
|
opcode = opcode.u.gdual->mod3[goffset];
|
|
|
|
|
else
|
|
|
|
|
opcode = opcode.u.gdual->mod012[goffset];
|
|
|
|
|
break;
|
|
|
|
|
case RMExt:
|
2011-06-01 12:34:25 +00:00
|
|
|
goffset = ctxt->modrm & 7;
|
2011-04-04 10:39:31 +00:00
|
|
|
opcode = opcode.u.group[goffset];
|
2011-04-24 11:09:59 +00:00
|
|
|
break;
|
|
|
|
|
case Prefix:
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->rep_prefix && op_prefix)
|
2011-07-30 09:03:34 +00:00
|
|
|
return EMULATION_FAILED;
|
2011-06-01 12:34:25 +00:00
|
|
|
simd_prefix = op_prefix ? 0x66 : ctxt->rep_prefix;
|
2011-04-24 11:09:59 +00:00
|
|
|
switch (simd_prefix) {
|
|
|
|
|
case 0x00: opcode = opcode.u.gprefix->pfx_no; break;
|
|
|
|
|
case 0x66: opcode = opcode.u.gprefix->pfx_66; break;
|
|
|
|
|
case 0xf2: opcode = opcode.u.gprefix->pfx_f2; break;
|
|
|
|
|
case 0xf3: opcode = opcode.u.gprefix->pfx_f3; break;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
default:
|
2011-07-30 09:03:34 +00:00
|
|
|
return EMULATION_FAILED;
|
2011-03-29 09:34:38 +00:00
|
|
|
}
|
2011-04-24 11:09:59 +00:00
|
|
|
|
2011-09-13 07:45:42 +00:00
|
|
|
ctxt->d &= ~(u64)GroupMask;
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->d |= opcode.flags;
|
2011-03-29 09:34:38 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->execute = opcode.u.execute;
|
|
|
|
|
ctxt->check_perm = opcode.check_perm;
|
|
|
|
|
ctxt->intercept = opcode.intercept;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
/* Unrecognised? */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->d == 0 || (ctxt->d & Undefined))
|
2011-07-30 09:03:34 +00:00
|
|
|
return EMULATION_FAILED;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (!(ctxt->d & VendorSpecific) && ctxt->only_vendor_specific_insn)
|
2011-07-30 09:03:34 +00:00
|
|
|
return EMULATION_FAILED;
|
2011-02-01 14:32:03 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (mode == X86EMUL_MODE_PROT64 && (ctxt->d & Stack))
|
|
|
|
|
ctxt->op_bytes = 8;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->d & Op3264) {
|
2010-08-01 11:46:54 +00:00
|
|
|
if (mode == X86EMUL_MODE_PROT64)
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes = 8;
|
2010-08-01 11:46:54 +00:00
|
|
|
else
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->op_bytes = 4;
|
2010-08-01 11:46:54 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->d & Sse)
|
|
|
|
|
ctxt->op_bytes = 16;
|
2012-04-09 15:40:02 +00:00
|
|
|
else if (ctxt->d & Mmx)
|
|
|
|
|
ctxt->op_bytes = 8;
|
2011-03-29 09:41:27 +00:00
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
/* ModRM and SIB bytes. */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->d & ModRM) {
|
2011-09-13 07:45:40 +00:00
|
|
|
rc = decode_modrm(ctxt, &ctxt->memop);
|
2011-06-01 12:34:25 +00:00
|
|
|
if (!ctxt->has_seg_override)
|
|
|
|
|
set_seg_override(ctxt, ctxt->modrm_seg);
|
|
|
|
|
} else if (ctxt->d & MemAbs)
|
2011-09-13 07:45:40 +00:00
|
|
|
rc = decode_abs(ctxt, &ctxt->memop);
|
2010-07-29 12:11:52 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (!ctxt->has_seg_override)
|
|
|
|
|
set_seg_override(ctxt, VCPU_SREG_DS);
|
2010-07-29 12:11:52 +00:00
|
|
|
|
2011-09-13 07:45:40 +00:00
|
|
|
ctxt->memop.addr.mem.seg = seg_override(ctxt);
|
2010-07-29 12:11:52 +00:00
|
|
|
|
2011-09-13 07:45:40 +00:00
|
|
|
if (ctxt->memop.type == OP_MEM && ctxt->ad_bytes != 8)
|
|
|
|
|
ctxt->memop.addr.mem.ea = (u32)ctxt->memop.addr.mem.ea;
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Decode and fetch the source operand: register, memory
|
|
|
|
|
* or immediate.
|
|
|
|
|
*/
|
2011-09-13 07:45:47 +00:00
|
|
|
rc = decode_operand(ctxt, &ctxt->src, (ctxt->d >> SrcShift) & OpMask);
|
2010-08-18 16:20:21 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
/*
|
|
|
|
|
* Decode and fetch the second source operand: register, memory
|
|
|
|
|
* or immediate.
|
|
|
|
|
*/
|
2011-09-13 07:45:43 +00:00
|
|
|
rc = decode_operand(ctxt, &ctxt->src2, (ctxt->d >> Src2Shift) & OpMask);
|
2010-08-18 16:20:21 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
|
2010-07-29 12:11:52 +00:00
|
|
|
/* Decode and fetch the destination operand: register or memory. */
|
2011-09-13 07:45:41 +00:00
|
|
|
rc = decode_operand(ctxt, &ctxt->dst, (ctxt->d >> DstShift) & OpMask);
|
2010-07-29 12:11:52 +00:00
|
|
|
|
|
|
|
|
done:
|
2011-09-13 07:45:40 +00:00
|
|
|
if (ctxt->memopp && ctxt->memopp->type == OP_MEM && ctxt->rip_relative)
|
|
|
|
|
ctxt->memopp->addr.mem.ea += ctxt->_eip;
|
2011-06-19 16:21:11 +00:00
|
|
|
|
2011-07-30 09:03:34 +00:00
|
|
|
return (rc != X86EMUL_CONTINUE) ? EMULATION_FAILED : EMULATION_OK;
|
2010-07-29 12:11:52 +00:00
|
|
|
}
|
|
|
|
|
|
2011-09-22 09:02:48 +00:00
|
|
|
bool x86_page_table_writing_insn(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
return ctxt->d & PageTable;
|
|
|
|
|
}
|
|
|
|
|
|
2010-08-25 09:47:42 +00:00
|
|
|
static bool string_insn_completed(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
/* The second termination condition only applies for REPE
|
|
|
|
|
* and REPNE. Test if the repeat string operation prefix is
|
|
|
|
|
* REPE/REPZ or REPNE/REPNZ and if it's the case it tests the
|
|
|
|
|
* corresponding termination condition according to:
|
|
|
|
|
* - if REPE/REPZ and ZF = 0 then done
|
|
|
|
|
* - if REPNE/REPNZ and ZF = 1 then done
|
|
|
|
|
*/
|
2011-06-01 12:34:25 +00:00
|
|
|
if (((ctxt->b == 0xa6) || (ctxt->b == 0xa7) ||
|
|
|
|
|
(ctxt->b == 0xae) || (ctxt->b == 0xaf))
|
|
|
|
|
&& (((ctxt->rep_prefix == REPE_PREFIX) &&
|
2010-08-25 09:47:42 +00:00
|
|
|
((ctxt->eflags & EFLG_ZF) == 0))
|
2011-06-01 12:34:25 +00:00
|
|
|
|| ((ctxt->rep_prefix == REPNE_PREFIX) &&
|
2010-08-25 09:47:42 +00:00
|
|
|
((ctxt->eflags & EFLG_ZF) == EFLG_ZF))))
|
|
|
|
|
return true;
|
|
|
|
|
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-09 15:40:02 +00:00
|
|
|
static int flush_pending_x87_faults(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
bool fault = false;
|
|
|
|
|
|
|
|
|
|
ctxt->ops->get_fpu(ctxt);
|
|
|
|
|
asm volatile("1: fwait \n\t"
|
|
|
|
|
"2: \n\t"
|
|
|
|
|
".pushsection .fixup,\"ax\" \n\t"
|
|
|
|
|
"3: \n\t"
|
|
|
|
|
"movb $1, %[fault] \n\t"
|
|
|
|
|
"jmp 2b \n\t"
|
|
|
|
|
".popsection \n\t"
|
|
|
|
|
_ASM_EXTABLE(1b, 3b)
|
2012-04-22 12:12:50 +00:00
|
|
|
: [fault]"+qm"(fault));
|
2012-04-09 15:40:02 +00:00
|
|
|
ctxt->ops->put_fpu(ctxt);
|
|
|
|
|
|
|
|
|
|
if (unlikely(fault))
|
|
|
|
|
return emulate_exception(ctxt, MF_VECTOR, 0, false);
|
|
|
|
|
|
|
|
|
|
return X86EMUL_CONTINUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void fetch_possible_mmx_operand(struct x86_emulate_ctxt *ctxt,
|
|
|
|
|
struct operand *op)
|
|
|
|
|
{
|
|
|
|
|
if (op->type == OP_MM)
|
|
|
|
|
read_mmx_reg(ctxt, &op->mm_val, op->addr.mm);
|
|
|
|
|
}
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
|
2011-05-14 16:00:52 +00:00
|
|
|
int x86_emulate_insn(struct x86_emulate_ctxt *ctxt)
|
2007-09-18 09:27:19 +00:00
|
|
|
{
|
2010-07-29 12:11:50 +00:00
|
|
|
struct x86_emulate_ops *ops = ctxt->ops;
|
2010-02-12 06:57:56 +00:00
|
|
|
int rc = X86EMUL_CONTINUE;
|
2011-06-01 12:34:25 +00:00
|
|
|
int saved_dst_type = ctxt->dst.type;
|
2007-09-18 09:27:19 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->mem_read.pos = 0;
|
2009-05-12 20:21:06 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->mode == X86EMUL_MODE_PROT64 && (ctxt->d & No64)) {
|
2010-11-22 15:53:25 +00:00
|
|
|
rc = emulate_ud(ctxt);
|
2010-02-11 12:43:14 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2010-02-10 12:21:36 +00:00
|
|
|
/* LOCK prefix is allowed only with some instructions */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->lock_prefix && (!(ctxt->d & Lock) || ctxt->dst.type != OP_MEM)) {
|
2010-11-22 15:53:25 +00:00
|
|
|
rc = emulate_ud(ctxt);
|
2010-02-10 12:21:36 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->d & SrcMask) == SrcMemFAddr && ctxt->src.type != OP_MEM) {
|
2010-11-22 15:53:25 +00:00
|
|
|
rc = emulate_ud(ctxt);
|
2010-08-26 08:06:15 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-09 15:40:02 +00:00
|
|
|
if (((ctxt->d & (Sse|Mmx)) && ((ops->get_cr(ctxt, 0) & X86_CR0_EM)))
|
|
|
|
|
|| ((ctxt->d & Sse) && !(ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR))) {
|
2011-03-29 09:41:27 +00:00
|
|
|
rc = emulate_ud(ctxt);
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-09 15:40:02 +00:00
|
|
|
if ((ctxt->d & (Sse|Mmx)) && (ops->get_cr(ctxt, 0) & X86_CR0_TS)) {
|
2011-03-29 09:41:27 +00:00
|
|
|
rc = emulate_nm(ctxt);
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2012-04-09 15:40:02 +00:00
|
|
|
if (ctxt->d & Mmx) {
|
|
|
|
|
rc = flush_pending_x87_faults(ctxt);
|
|
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
/*
|
|
|
|
|
* Now that we know the fpu is exception safe, we can fetch
|
|
|
|
|
* operands from it.
|
|
|
|
|
*/
|
|
|
|
|
fetch_possible_mmx_operand(ctxt, &ctxt->src);
|
|
|
|
|
fetch_possible_mmx_operand(ctxt, &ctxt->src2);
|
|
|
|
|
if (!(ctxt->d & Mov))
|
|
|
|
|
fetch_possible_mmx_operand(ctxt, &ctxt->dst);
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (unlikely(ctxt->guest_mode) && ctxt->intercept) {
|
|
|
|
|
rc = emulator_check_intercept(ctxt, ctxt->intercept,
|
2011-04-04 10:39:27 +00:00
|
|
|
X86_ICPT_PRE_EXCEPT);
|
2011-04-04 10:39:22 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2010-02-10 12:21:35 +00:00
|
|
|
/* Privileged instruction can be executed only in CPL=0 */
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->d & Priv) && ops->cpl(ctxt)) {
|
2010-11-22 15:53:25 +00:00
|
|
|
rc = emulate_gp(ctxt, 0);
|
2010-02-10 12:21:35 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:26 +00:00
|
|
|
/* Instruction can only be executed in protected mode */
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->d & Prot) && !(ctxt->mode & X86EMUL_MODE_PROT)) {
|
2011-04-04 10:39:26 +00:00
|
|
|
rc = emulate_ud(ctxt);
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-04-04 10:39:25 +00:00
|
|
|
/* Do instruction specific permission checks */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->check_perm) {
|
|
|
|
|
rc = ctxt->check_perm(ctxt);
|
2011-04-04 10:39:25 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (unlikely(ctxt->guest_mode) && ctxt->intercept) {
|
|
|
|
|
rc = emulator_check_intercept(ctxt, ctxt->intercept,
|
2011-04-04 10:39:27 +00:00
|
|
|
X86_ICPT_POST_EXCEPT);
|
2011-04-04 10:39:22 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->rep_prefix && (ctxt->d & String)) {
|
2007-11-27 17:05:37 +00:00
|
|
|
/* All REP prefixes have the same first termination condition */
|
2012-08-27 20:46:17 +00:00
|
|
|
if (address_mask(ctxt, reg_read(ctxt, VCPU_REGS_RCX)) == 0) {
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->eip = ctxt->_eip;
|
2007-11-27 17:05:37 +00:00
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->src.type == OP_MEM) && !(ctxt->d & NoAccess)) {
|
|
|
|
|
rc = segmented_read(ctxt, ctxt->src.addr.mem,
|
|
|
|
|
ctxt->src.valptr, ctxt->src.bytes);
|
2010-01-20 07:47:21 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
2007-09-18 09:27:19 +00:00
|
|
|
goto done;
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.orig_val64 = ctxt->src.val64;
|
2007-09-18 09:27:19 +00:00
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->src2.type == OP_MEM) {
|
|
|
|
|
rc = segmented_read(ctxt, ctxt->src2.addr.mem,
|
|
|
|
|
&ctxt->src2.val, ctxt->src2.bytes);
|
2010-02-25 14:36:42 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->d & DstMask) == ImplicitOps)
|
2007-09-18 09:27:19 +00:00
|
|
|
goto special_insn;
|
|
|
|
|
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->dst.type == OP_MEM) && !(ctxt->d & Mov)) {
|
2010-03-18 13:20:20 +00:00
|
|
|
/* optimisation - avoid slow emulated read if Mov */
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = segmented_read(ctxt, ctxt->dst.addr.mem,
|
|
|
|
|
&ctxt->dst.val, ctxt->dst.bytes);
|
2010-03-18 13:20:20 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
2007-01-23 04:40:40 +00:00
|
|
|
}
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.orig_val = ctxt->dst.val;
|
2007-01-23 04:40:40 +00:00
|
|
|
|
2007-11-27 17:30:56 +00:00
|
|
|
special_insn:
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (unlikely(ctxt->guest_mode) && ctxt->intercept) {
|
|
|
|
|
rc = emulator_check_intercept(ctxt, ctxt->intercept,
|
2011-04-04 10:39:27 +00:00
|
|
|
X86_ICPT_POST_MEMACCESS);
|
2011-04-04 10:39:22 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->execute) {
|
|
|
|
|
rc = ctxt->execute(ctxt);
|
2010-07-29 12:11:51 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
goto writeback;
|
|
|
|
|
}
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->twobyte)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
goto twobyte_insn;
|
|
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->b) {
|
2007-10-31 09:15:56 +00:00
|
|
|
case 0x40 ... 0x47: /* inc r16/r32 */
|
2011-09-07 13:41:38 +00:00
|
|
|
emulate_1op(ctxt, "inc");
|
2007-10-31 09:15:56 +00:00
|
|
|
break;
|
|
|
|
|
case 0x48 ... 0x4f: /* dec r16/r32 */
|
2011-09-07 13:41:38 +00:00
|
|
|
emulate_1op(ctxt, "dec");
|
2007-10-31 09:15:56 +00:00
|
|
|
break;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
case 0x63: /* movsxd */
|
2007-09-18 09:27:19 +00:00
|
|
|
if (ctxt->mode != X86EMUL_MODE_PROT64)
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
goto cannot_emulate;
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = (s32) ctxt->src.val;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2009-04-12 10:36:30 +00:00
|
|
|
case 0x70 ... 0x7f: /* jcc (short) */
|
2011-06-01 12:34:25 +00:00
|
|
|
if (test_cc(ctxt->b, ctxt->eflags))
|
|
|
|
|
jmp_rel(ctxt, ctxt->src.val);
|
2007-11-27 17:30:56 +00:00
|
|
|
break;
|
2007-09-15 07:35:36 +00:00
|
|
|
case 0x8d: /* lea r16/r32, m */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ctxt->src.addr.mem.ea;
|
2007-09-15 07:35:36 +00:00
|
|
|
break;
|
2010-08-01 09:41:59 +00:00
|
|
|
case 0x90 ... 0x97: /* nop / xchg reg, rax */
|
2012-08-27 20:46:17 +00:00
|
|
|
if (ctxt->dst.addr.reg == reg_rmw(ctxt, VCPU_REGS_RAX))
|
2010-08-04 11:41:04 +00:00
|
|
|
break;
|
2011-05-29 12:59:09 +00:00
|
|
|
rc = em_xchg(ctxt);
|
|
|
|
|
break;
|
2010-08-18 08:43:13 +00:00
|
|
|
case 0x98: /* cbw/cwde/cdqe */
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->op_bytes) {
|
|
|
|
|
case 2: ctxt->dst.val = (s8)ctxt->dst.val; break;
|
|
|
|
|
case 4: ctxt->dst.val = (s16)ctxt->dst.val; break;
|
|
|
|
|
case 8: ctxt->dst.val = (s32)ctxt->dst.val; break;
|
2010-08-18 08:43:13 +00:00
|
|
|
}
|
|
|
|
|
break;
|
2007-11-27 17:30:56 +00:00
|
|
|
case 0xc0 ... 0xc1:
|
2011-05-01 17:29:17 +00:00
|
|
|
rc = em_grp2(ctxt);
|
2007-11-27 17:30:56 +00:00
|
|
|
break;
|
2010-08-04 11:38:06 +00:00
|
|
|
case 0xcc: /* int3 */
|
2011-05-29 13:02:55 +00:00
|
|
|
rc = emulate_int(ctxt, 3);
|
|
|
|
|
break;
|
2010-08-04 11:38:06 +00:00
|
|
|
case 0xcd: /* int n */
|
2011-06-01 12:34:25 +00:00
|
|
|
rc = emulate_int(ctxt, ctxt->src.val);
|
2010-08-04 11:38:06 +00:00
|
|
|
break;
|
|
|
|
|
case 0xce: /* into */
|
2011-05-29 13:02:55 +00:00
|
|
|
if (ctxt->eflags & EFLG_OF)
|
|
|
|
|
rc = emulate_int(ctxt, 4);
|
2010-08-04 11:38:06 +00:00
|
|
|
break;
|
2007-11-27 17:30:56 +00:00
|
|
|
case 0xd0 ... 0xd1: /* Grp2 */
|
2011-05-01 17:29:17 +00:00
|
|
|
rc = em_grp2(ctxt);
|
2007-11-27 17:30:56 +00:00
|
|
|
break;
|
|
|
|
|
case 0xd2 ... 0xd3: /* Grp2 */
|
2012-08-27 20:46:17 +00:00
|
|
|
ctxt->src.val = reg_read(ctxt, VCPU_REGS_RCX);
|
2011-05-01 17:29:17 +00:00
|
|
|
rc = em_grp2(ctxt);
|
2007-11-27 17:30:56 +00:00
|
|
|
break;
|
2007-09-18 23:34:25 +00:00
|
|
|
case 0xe9: /* jmp rel */
|
2011-05-29 12:56:26 +00:00
|
|
|
case 0xeb: /* jmp rel short */
|
2011-06-01 12:34:25 +00:00
|
|
|
jmp_rel(ctxt, ctxt->src.val);
|
|
|
|
|
ctxt->dst.type = OP_NONE; /* Disable writeback. */
|
2007-09-18 23:34:25 +00:00
|
|
|
break;
|
2007-11-27 17:14:21 +00:00
|
|
|
case 0xf4: /* hlt */
|
2011-04-20 12:43:05 +00:00
|
|
|
ctxt->ops->halt(ctxt);
|
2008-07-06 13:51:26 +00:00
|
|
|
break;
|
2007-11-27 17:14:21 +00:00
|
|
|
case 0xf5: /* cmc */
|
|
|
|
|
/* complement carry flag from eflags reg */
|
|
|
|
|
ctxt->eflags ^= EFLG_CF;
|
|
|
|
|
break;
|
|
|
|
|
case 0xf8: /* clc */
|
|
|
|
|
ctxt->eflags &= ~EFLG_CF;
|
|
|
|
|
break;
|
2010-08-05 12:42:49 +00:00
|
|
|
case 0xf9: /* stc */
|
|
|
|
|
ctxt->eflags |= EFLG_CF;
|
|
|
|
|
break;
|
2008-09-01 01:52:24 +00:00
|
|
|
case 0xfc: /* cld */
|
|
|
|
|
ctxt->eflags &= ~EFLG_DF;
|
|
|
|
|
break;
|
|
|
|
|
case 0xfd: /* std */
|
|
|
|
|
ctxt->eflags |= EFLG_DF;
|
|
|
|
|
break;
|
2010-07-25 11:51:16 +00:00
|
|
|
default:
|
|
|
|
|
goto cannot_emulate;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
}
|
2007-11-27 17:30:56 +00:00
|
|
|
|
2010-08-30 14:12:28 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
|
2007-11-27 17:30:56 +00:00
|
|
|
writeback:
|
2011-05-01 17:26:23 +00:00
|
|
|
rc = writeback(ctxt);
|
2010-02-12 06:57:56 +00:00
|
|
|
if (rc != X86EMUL_CONTINUE)
|
2007-11-27 17:30:56 +00:00
|
|
|
goto done;
|
|
|
|
|
|
2010-03-18 13:20:26 +00:00
|
|
|
/*
|
|
|
|
|
* restore dst type in case the decoding will be reused
|
|
|
|
|
* (happens for string instruction )
|
|
|
|
|
*/
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.type = saved_dst_type;
|
2010-03-18 13:20:26 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->d & SrcMask) == SrcSI)
|
|
|
|
|
string_addr_inc(ctxt, seg_override(ctxt),
|
|
|
|
|
VCPU_REGS_RSI, &ctxt->src);
|
2010-03-18 13:20:21 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if ((ctxt->d & DstMask) == DstDI)
|
2010-11-17 13:28:21 +00:00
|
|
|
string_addr_inc(ctxt, VCPU_SREG_ES, VCPU_REGS_RDI,
|
2011-06-01 12:34:25 +00:00
|
|
|
&ctxt->dst);
|
2010-03-18 13:20:22 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
if (ctxt->rep_prefix && (ctxt->d & String)) {
|
|
|
|
|
struct read_cache *r = &ctxt->io_read;
|
2012-08-27 20:46:17 +00:00
|
|
|
register_address_increment(ctxt, reg_rmw(ctxt, VCPU_REGS_RCX), -1);
|
2010-08-25 09:47:42 +00:00
|
|
|
|
2010-08-25 09:47:43 +00:00
|
|
|
if (!string_insn_completed(ctxt)) {
|
|
|
|
|
/*
|
|
|
|
|
* Re-enter guest when pio read ahead buffer is empty
|
|
|
|
|
* or, if it is not used, after each 1024 iteration.
|
|
|
|
|
*/
|
2012-08-27 20:46:17 +00:00
|
|
|
if ((r->end != 0 || reg_read(ctxt, VCPU_REGS_RCX) & 0x3ff) &&
|
2010-08-25 09:47:43 +00:00
|
|
|
(r->end == 0 || r->end != r->pos)) {
|
|
|
|
|
/*
|
|
|
|
|
* Reset read cache. Usually happens before
|
|
|
|
|
* decode, but since instruction is restarted
|
|
|
|
|
* we have to do it here.
|
|
|
|
|
*/
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->mem_read.end = 0;
|
2012-08-27 20:46:17 +00:00
|
|
|
writeback_registers(ctxt);
|
2010-08-25 09:47:43 +00:00
|
|
|
return EMULATION_RESTART;
|
|
|
|
|
}
|
|
|
|
|
goto done; /* skip rip writeback */
|
2010-08-17 08:22:17 +00:00
|
|
|
}
|
2010-03-18 13:20:26 +00:00
|
|
|
}
|
2010-08-25 09:47:43 +00:00
|
|
|
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->eip = ctxt->_eip;
|
2007-11-27 17:30:56 +00:00
|
|
|
|
|
|
|
|
done:
|
2010-11-22 15:53:21 +00:00
|
|
|
if (rc == X86EMUL_PROPAGATE_FAULT)
|
|
|
|
|
ctxt->have_exception = true;
|
2011-04-04 10:39:24 +00:00
|
|
|
if (rc == X86EMUL_INTERCEPTED)
|
|
|
|
|
return EMULATION_INTERCEPTED;
|
|
|
|
|
|
2012-08-27 20:46:17 +00:00
|
|
|
if (rc == X86EMUL_CONTINUE)
|
|
|
|
|
writeback_registers(ctxt);
|
|
|
|
|
|
2010-08-25 09:47:43 +00:00
|
|
|
return (rc == X86EMUL_UNHANDLEABLE) ? EMULATION_FAILED : EMULATION_OK;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
|
|
|
|
|
twobyte_insn:
|
2011-06-01 12:34:25 +00:00
|
|
|
switch (ctxt->b) {
|
2007-11-27 17:30:56 +00:00
|
|
|
case 0x09: /* wbinvd */
|
2011-04-21 19:16:05 +00:00
|
|
|
(ctxt->ops->wbinvd)(ctxt);
|
2010-06-30 04:25:15 +00:00
|
|
|
break;
|
|
|
|
|
case 0x08: /* invd */
|
2007-11-27 17:30:56 +00:00
|
|
|
case 0x0d: /* GrpP (prefetch) */
|
|
|
|
|
case 0x18: /* Grp16 (prefetch/nop) */
|
|
|
|
|
break;
|
|
|
|
|
case 0x20: /* mov cr, reg */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ops->get_cr(ctxt, ctxt->modrm_reg);
|
2007-11-27 17:30:56 +00:00
|
|
|
break;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
case 0x21: /* mov from dr to reg */
|
2011-06-01 12:34:25 +00:00
|
|
|
ops->get_dr(ctxt, ctxt->modrm_reg, &ctxt->dst.val);
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
|
|
|
|
case 0x40 ... 0x4f: /* cmov */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = ctxt->dst.orig_val = ctxt->src.val;
|
|
|
|
|
if (!test_cc(ctxt->b, ctxt->eflags))
|
|
|
|
|
ctxt->dst.type = OP_NONE; /* no writeback */
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2009-04-12 10:36:30 +00:00
|
|
|
case 0x80 ... 0x8f: /* jnz rel, etc*/
|
2011-06-01 12:34:25 +00:00
|
|
|
if (test_cc(ctxt->b, ctxt->eflags))
|
|
|
|
|
jmp_rel(ctxt, ctxt->src.val);
|
2007-11-27 17:30:56 +00:00
|
|
|
break;
|
2010-08-06 09:10:07 +00:00
|
|
|
case 0x90 ... 0x9f: /* setcc r/m8 */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.val = test_cc(ctxt->b, ctxt->eflags);
|
2010-08-06 09:10:07 +00:00
|
|
|
break;
|
2008-12-04 13:30:13 +00:00
|
|
|
case 0xa4: /* shld imm8, r, r/m */
|
|
|
|
|
case 0xa5: /* shld cl, r, r/m */
|
2011-09-07 13:41:36 +00:00
|
|
|
emulate_2op_cl(ctxt, "shld");
|
2008-12-04 13:30:13 +00:00
|
|
|
break;
|
|
|
|
|
case 0xac: /* shrd imm8, r, r/m */
|
|
|
|
|
case 0xad: /* shrd cl, r, r/m */
|
2011-09-07 13:41:36 +00:00
|
|
|
emulate_2op_cl(ctxt, "shrd");
|
2008-12-04 13:30:13 +00:00
|
|
|
break;
|
2008-07-10 20:08:15 +00:00
|
|
|
case 0xae: /* clflush */
|
|
|
|
|
break;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
case 0xb6 ... 0xb7: /* movzx */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.bytes = ctxt->op_bytes;
|
2012-06-11 16:40:15 +00:00
|
|
|
ctxt->dst.val = (ctxt->src.bytes == 1) ? (u8) ctxt->src.val
|
2011-06-01 12:34:25 +00:00
|
|
|
: (u16) ctxt->src.val;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
|
|
|
|
case 0xbe ... 0xbf: /* movsx */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.bytes = ctxt->op_bytes;
|
2012-06-11 16:40:15 +00:00
|
|
|
ctxt->dst.val = (ctxt->src.bytes == 1) ? (s8) ctxt->src.val :
|
2011-06-01 12:34:25 +00:00
|
|
|
(s16) ctxt->src.val;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
break;
|
2010-08-17 01:19:34 +00:00
|
|
|
case 0xc0 ... 0xc1: /* xadd */
|
2011-09-07 13:41:35 +00:00
|
|
|
emulate_2op_SrcV(ctxt, "add");
|
2010-08-17 01:19:34 +00:00
|
|
|
/* Write back the register source. */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->src.val = ctxt->dst.orig_val;
|
|
|
|
|
write_register_operand(&ctxt->src);
|
2010-08-17 01:19:34 +00:00
|
|
|
break;
|
2007-10-15 06:24:20 +00:00
|
|
|
case 0xc3: /* movnti */
|
2011-06-01 12:34:25 +00:00
|
|
|
ctxt->dst.bytes = ctxt->op_bytes;
|
|
|
|
|
ctxt->dst.val = (ctxt->op_bytes == 4) ? (u32) ctxt->src.val :
|
|
|
|
|
(u64) ctxt->src.val;
|
2007-10-15 06:24:20 +00:00
|
|
|
break;
|
2010-07-25 11:51:16 +00:00
|
|
|
default:
|
|
|
|
|
goto cannot_emulate;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
}
|
2010-08-30 14:12:28 +00:00
|
|
|
|
|
|
|
|
if (rc != X86EMUL_CONTINUE)
|
|
|
|
|
goto done;
|
|
|
|
|
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
goto writeback;
|
|
|
|
|
|
|
|
|
|
cannot_emulate:
|
2011-03-28 14:57:49 +00:00
|
|
|
return EMULATION_FAILED;
|
[PATCH] kvm: userspace interface
web site: http://kvm.sourceforge.net
mailing list: kvm-devel@lists.sourceforge.net
(http://lists.sourceforge.net/lists/listinfo/kvm-devel)
The following patchset adds a driver for Intel's hardware virtualization
extensions to the x86 architecture. The driver adds a character device
(/dev/kvm) that exposes the virtualization capabilities to userspace. Using
this driver, a process can run a virtual machine (a "guest") in a fully
virtualized PC containing its own virtual hard disks, network adapters, and
display.
Using this driver, one can start multiple virtual machines on a host.
Each virtual machine is a process on the host; a virtual cpu is a thread in
that process. kill(1), nice(1), top(1) work as expected. In effect, the
driver adds a third execution mode to the existing two: we now have kernel
mode, user mode, and guest mode. Guest mode has its own address space mapping
guest physical memory (which is accessible to user mode by mmap()ing
/dev/kvm). Guest mode has no access to any I/O devices; any such access is
intercepted and directed to user mode for emulation.
The driver supports i386 and x86_64 hosts and guests. All combinations are
allowed except x86_64 guest on i386 host. For i386 guests and hosts, both pae
and non-pae paging modes are supported.
SMP hosts and UP guests are supported. At the moment only Intel
hardware is supported, but AMD virtualization support is being worked on.
Performance currently is non-stellar due to the naive implementation of the
mmu virtualization, which throws away most of the shadow page table entries
every context switch. We plan to address this in two ways:
- cache shadow page tables across tlb flushes
- wait until AMD and Intel release processors with nested page tables
Currently a virtual desktop is responsive but consumes a lot of CPU. Under
Windows I tried playing pinball and watching a few flash movies; with a recent
CPU one can hardly feel the virtualization. Linux/X is slower, probably due
to X being in a separate process.
In addition to the driver, you need a slightly modified qemu to provide I/O
device emulation and the BIOS.
Caveats (akpm: might no longer be true):
- The Windows install currently bluescreens due to a problem with the
virtual APIC. We are working on a fix. A temporary workaround is to
use an existing image or install through qemu
- Windows 64-bit does not work. That's also true for qemu, so it's
probably a problem with the device model.
[bero@arklinux.org: build fix]
[simon.kagstrom@bth.se: build fix, other fixes]
[uril@qumranet.com: KVM: Expose interrupt bitmap]
[akpm@osdl.org: i386 build fix]
[mingo@elte.hu: i386 fixes]
[rdreier@cisco.com: add log levels to all printks]
[randy.dunlap@oracle.com: Fix sparse NULL and C99 struct init warnings]
[anthony@codemonkey.ws: KVM: AMD SVM: 32-bit host support]
Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Simon Kagstrom <simon.kagstrom@bth.se>
Cc: Bernhard Rosenkraenzer <bero@arklinux.org>
Signed-off-by: Uri Lublin <uril@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 10:21:36 +00:00
|
|
|
}
|
2012-08-27 20:46:17 +00:00
|
|
|
|
|
|
|
|
void emulator_invalidate_register_cache(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
invalidate_registers(ctxt);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void emulator_writeback_register_cache(struct x86_emulate_ctxt *ctxt)
|
|
|
|
|
{
|
|
|
|
|
writeback_registers(ctxt);
|
|
|
|
|
}
|
