qemu-tech.texi 12.3 KB
Newer Older
1 2
@node Implementation notes
@appendix Implementation notes
3 4

@menu
5 6 7 8
* CPU emulation::
* Translator Internals::
* QEMU compared to other emulators::
* Bibliography::
9 10
@end menu

11
@node CPU emulation
12
@section CPU emulation
bellard's avatar
bellard committed
13

14
@menu
15 16 17 18 19 20
* x86::     x86 and x86-64 emulation
* ARM::     ARM emulation
* MIPS::    MIPS emulation
* PPC::     PowerPC emulation
* SPARC::   Sparc32 and Sparc64 emulation
* Xtensa::  Xtensa emulation
21 22
@end menu

23
@node x86
24
@subsection x86 and x86-64 emulation
bellard's avatar
bellard committed
25 26 27

QEMU x86 target features:

28
@itemize
bellard's avatar
bellard committed
29

30
@item The virtual x86 CPU supports 16 bit and 32 bit addressing with segmentation.
31 32 33
LDT/GDT and IDT are emulated. VM86 mode is also supported to run
DOSEMU. There is some support for MMX/3DNow!, SSE, SSE2, SSE3, SSSE3,
and SSE4 as well as x86-64 SVM.
bellard's avatar
bellard committed
34 35 36 37 38

@item Support of host page sizes bigger than 4KB in user mode emulation.

@item QEMU can emulate itself on x86.

39
@item An extensive Linux x86 CPU test program is included @file{tests/test-i386}.
bellard's avatar
bellard committed
40 41 42 43 44 45
It can be used to test other x86 virtual CPUs.

@end itemize

Current QEMU limitations:

46
@itemize
bellard's avatar
bellard committed
47

48
@item Limited x86-64 support.
bellard's avatar
bellard committed
49 50 51

@item IPC syscalls are missing.

52
@item The x86 segment limits and access rights are not tested at every
bellard's avatar
bellard committed
53 54 55 56 57
memory access (yet). Hopefully, very few OSes seem to rely on that for
normal use.

@end itemize

58
@node ARM
59
@subsection ARM emulation
bellard's avatar
bellard committed
60 61 62 63 64 65 66 67 68 69 70

@itemize

@item Full ARM 7 user emulation.

@item NWFPE FPU support included in user Linux emulation.

@item Can run most ARM Linux binaries.

@end itemize

71
@node MIPS
72
@subsection MIPS emulation
ths's avatar
ths committed
73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93

@itemize

@item The system emulation allows full MIPS32/MIPS64 Release 2 emulation,
including privileged instructions, FPU and MMU, in both little and big
endian modes.

@item The Linux userland emulation can run many 32 bit MIPS Linux binaries.

@end itemize

Current QEMU limitations:

@itemize

@item Self-modifying code is not always handled correctly.

@item 64 bit userland emulation is not implemented.

@item The system emulation is not complete enough to run real firmware.

94 95
@item The watchpoint debug facility is not implemented.

ths's avatar
ths committed
96 97
@end itemize

98
@node PPC
99
@subsection PowerPC emulation
bellard's avatar
bellard committed
100 101 102

@itemize

103
@item Full PowerPC 32 bit emulation, including privileged instructions,
bellard's avatar
bellard committed
104 105 106 107 108 109
FPU and MMU.

@item Can run most PowerPC Linux binaries.

@end itemize

110
@node SPARC
111
@subsection Sparc32 and Sparc64 emulation
bellard's avatar
bellard committed
112 113 114

@itemize

blueswir1's avatar
blueswir1 committed
115
@item Full SPARC V8 emulation, including privileged
bellard's avatar
bellard committed
116
instructions, FPU and MMU. SPARC V9 emulation includes most privileged
117
and VIS instructions, FPU and I/D MMU. Alignment is fully enforced.
bellard's avatar
bellard committed
118

119 120
@item Can run most 32-bit SPARC Linux binaries, SPARC32PLUS Linux binaries and
some 64-bit SPARC Linux binaries.
bellard's avatar
bellard committed
121 122 123 124 125

@end itemize

Current QEMU limitations:

126
@itemize
bellard's avatar
bellard committed
127 128 129

@item IPC syscalls are missing.

130
@item Floating point exception support is buggy.
bellard's avatar
bellard committed
131 132 133

@item Atomic instructions are not correctly implemented.

134 135 136 137
@item There are still some problems with Sparc64 emulators.

@end itemize

138
@node Xtensa
139
@subsection Xtensa emulation
140 141 142 143 144

@itemize

@item Core Xtensa ISA emulation, including most options: code density,
loop, extended L32R, 16- and 32-bit multiplication, 32-bit division,
145 146
MAC16, miscellaneous operations, boolean, FP coprocessor, coprocessor
context, debug, multiprocessor synchronization,
147 148 149 150 151
conditional store, exceptions, relocatable vectors, unaligned exception,
interrupts (including high priority and timer), hardware alignment,
region protection, region translation, MMU, windowed registers, thread
pointer, processor ID.

152 153 154
@item Not implemented options: data/instruction cache (including cache
prefetch and locking), XLMI, processor interface. Also options not
covered by the core ISA (e.g. FLIX, wide branches) are not implemented.
155 156 157 158 159 160 161 162

@item Can run most Xtensa Linux binaries.

@item New core configuration that requires no additional instructions
may be created from overlay with minimal amount of hand-written code.

@end itemize

163
@node Translator Internals
164
@section Translator Internals
bellard's avatar
bellard committed
165 166 167 168 169 170 171

QEMU is a dynamic translator. When it first encounters a piece of code,
it converts it to the host instruction set. Usually dynamic translators
are very complicated and highly CPU dependent. QEMU uses some tricks
which make it relatively easily portable and simple while achieving good
performances.

172 173
QEMU's dynamic translation backend is called TCG, for "Tiny Code
Generator". For more information, please take a look at @code{tcg/README}.
bellard's avatar
bellard committed
174

175
Some notable features of QEMU's dynamic translator are:
bellard's avatar
bellard committed
176

177 178 179
@table @strong

@item CPU state optimisations:
180 181 182 183 184 185
The target CPUs have many internal states which change the way it
evaluates instructions. In order to achieve a good speed, the
translation phase considers that some state information of the virtual
CPU cannot change in it. The state is recorded in the Translation
Block (TB). If the state changes (e.g. privilege level), a new TB will
be generated and the previous TB won't be used anymore until the state
186 187
matches the state recorded in the previous TB. The same idea can be applied
to other aspects of the CPU state.  For example, on x86, if the SS,
188 189
DS and ES segments have a zero base, then the translator does not even
generate an addition for the segment base.
bellard's avatar
bellard committed
190

191
@item Direct block chaining:
bellard's avatar
bellard committed
192
After each translated basic block is executed, QEMU uses the simulated
Gonglei's avatar
Gonglei committed
193
Program Counter (PC) and other cpu state information (such as the CS
bellard's avatar
bellard committed
194 195 196 197 198 199 200 201 202 203 204
segment base value) to find the next basic block.

In order to accelerate the most common cases where the new simulated PC
is known, QEMU can patch a basic block so that it jumps directly to the
next one.

The most portable code uses an indirect jump. An indirect jump makes
it easier to make the jump target modification atomic. On some host
architectures (such as x86 or PowerPC), the @code{JUMP} opcode is
directly patched so that the block chaining has no overhead.

205
@item Self-modifying code and translated code invalidation:
bellard's avatar
bellard committed
206 207 208 209
Self-modifying code is a special challenge in x86 emulation because no
instruction cache invalidation is signaled by the application when code
is modified.

210 211 212 213 214 215
User-mode emulation marks a host page as write-protected (if it is
not already read-only) every time translated code is generated for a
basic block.  Then, if a write access is done to the page, Linux raises
a SEGV signal. QEMU then invalidates all the translated code in the page
and enables write accesses to the page.  For system emulation, write
protection is achieved through the software MMU.
bellard's avatar
bellard committed
216 217 218

Correct translated code invalidation is done efficiently by maintaining
a linked list of every translated block contained in a given page. Other
219
linked lists are also maintained to undo direct block chaining.
bellard's avatar
bellard committed
220

221 222 223 224 225
On RISC targets, correctly written software uses memory barriers and
cache flushes, so some of the protection above would not be
necessary. However, QEMU still requires that the generated code always
matches the target instructions in memory in order to handle
exceptions correctly.
bellard's avatar
bellard committed
226

227
@item Exception support:
bellard's avatar
bellard committed
228
longjmp() is used when an exception such as division by zero is
229
encountered.
bellard's avatar
bellard committed
230 231

The host SIGSEGV and SIGBUS signal handlers are used to get invalid
232 233 234 235 236 237 238 239 240 241 242 243 244 245
memory accesses.  QEMU keeps a map from host program counter to
target program counter, and looks up where the exception happened
based on the host program counter at the exception point.

On some targets, some bits of the virtual CPU's state are not flushed to the
memory until the end of the translation block.  This is done for internal
emulation state that is rarely accessed directly by the program and/or changes
very often throughout the execution of a translation block---this includes
condition codes on x86, delay slots on SPARC, conditional execution on
ARM, and so on.  This state is stored for each target instruction, and
looked up on exceptions.

@item MMU emulation:
For system emulation QEMU uses a software MMU. In that mode, the MMU
246
virtual to physical address translation is done at every memory
247
access.
bellard's avatar
bellard committed
248

249
QEMU uses an address translation cache (TLB) to speed up the translation.
bellard's avatar
bellard committed
250
In order to avoid flushing the translated code each time the MMU
251
mappings change, all caches in QEMU are physically indexed.  This
252
means that each basic block is indexed with its physical address.
bellard's avatar
bellard committed
253

254 255 256 257 258 259 260 261 262 263 264
In order to avoid invalidating the basic block chain when MMU mappings
change, chaining is only performed when the destination of the jump
shares a page with the basic block that is performing the jump.

The MMU can also distinguish RAM and ROM memory areas from MMIO memory
areas.  Access is faster for RAM and ROM because the translation cache also
hosts the offset between guest address and host memory.  Accessing MMIO
memory areas instead calls out to C code for device emulation.
Finally, the MMU helps tracking dirty pages and pages pointed to by
translation blocks.
@end table
265

266
@node QEMU compared to other emulators
267
@section QEMU compared to other emulators
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316

Like bochs [1], QEMU emulates an x86 CPU. But QEMU is much faster than
bochs as it uses dynamic compilation. Bochs is closely tied to x86 PC
emulation while QEMU can emulate several processors.

Like Valgrind [2], QEMU does user space emulation and dynamic
translation. Valgrind is mainly a memory debugger while QEMU has no
support for it (QEMU could be used to detect out of bound memory
accesses as Valgrind, but it has no support to track uninitialised data
as Valgrind does). The Valgrind dynamic translator generates better code
than QEMU (in particular it does register allocation) but it is closely
tied to an x86 host and target and has no support for precise exceptions
and system emulation.

EM86 [3] is the closest project to user space QEMU (and QEMU still uses
some of its code, in particular the ELF file loader). EM86 was limited
to an alpha host and used a proprietary and slow interpreter (the
interpreter part of the FX!32 Digital Win32 code translator [4]).

TWIN from Willows Software was a Windows API emulator like Wine. It is less
accurate than Wine but includes a protected mode x86 interpreter to launch
x86 Windows executables. Such an approach has greater potential because most
of the Windows API is executed natively but it is far more difficult to
develop because all the data structures and function parameters exchanged
between the API and the x86 code must be converted.

User mode Linux [5] was the only solution before QEMU to launch a
Linux kernel as a process while not needing any host kernel
patches. However, user mode Linux requires heavy kernel patches while
QEMU accepts unpatched Linux kernels. The price to pay is that QEMU is
slower.

The Plex86 [6] PC virtualizer is done in the same spirit as the now
obsolete qemu-fast system emulator. It requires a patched Linux kernel
to work (you cannot launch the same kernel on your PC), but the
patches are really small. As it is a PC virtualizer (no emulation is
done except for some privileged instructions), it has the potential of
being faster than QEMU. The downside is that a complicated (and
potentially unsafe) host kernel patch is needed.

The commercial PC Virtualizers (VMWare [7], VirtualPC [8]) are faster
than QEMU (without virtualization), but they all need specific, proprietary
and potentially unsafe host drivers. Moreover, they are unable to
provide cycle exact simulation as an emulator can.

VirtualBox [9], Xen [10] and KVM [11] are based on QEMU. QEMU-SystemC
[12] uses QEMU to simulate a system where some hardware devices are
developed in SystemC.

317
@node Bibliography
318
@section Bibliography
bellard's avatar
bellard committed
319 320 321

@table @asis

322
@item [1]
323 324
@url{http://bochs.sourceforge.net/}, the Bochs IA-32 Emulator Project,
by Kevin Lawton et al.
bellard's avatar
bellard committed
325 326

@item [2]
327 328
@url{http://www.valgrind.org/}, Valgrind, an open-source memory debugger
for GNU/Linux.
bellard's avatar
bellard committed
329 330

@item [3]
331 332
@url{http://ftp.dreamtime.org/pub/linux/Linux-Alpha/em86/v0.2/docs/em86.html},
the EM86 x86 emulator on Alpha-Linux.
bellard's avatar
bellard committed
333 334

@item [4]
335
@url{http://www.usenix.org/publications/library/proceedings/usenix-nt97/@/full_papers/chernoff/chernoff.pdf},
bellard's avatar
bellard committed
336 337 338
DIGITAL FX!32: Running 32-Bit x86 Applications on Alpha NT, by Anton
Chernoff and Ray Hookway.

339
@item [5]
340
@url{http://user-mode-linux.sourceforge.net/},
bellard's avatar
bellard committed
341 342
The User-mode Linux Kernel.

343
@item [6]
344
@url{http://www.plex86.org/},
bellard's avatar
bellard committed
345 346
The new Plex86 project.

347
@item [7]
348
@url{http://www.vmware.com/},
bellard's avatar
bellard committed
349 350
The VMWare PC virtualizer.

351 352
@item [8]
@url{https://www.microsoft.com/download/details.aspx?id=3702},
bellard's avatar
bellard committed
353 354
The VirtualPC PC virtualizer.

355
@item [9]
356 357 358
@url{http://virtualbox.org/},
The VirtualBox PC virtualizer.

359
@item [10]
360 361 362
@url{http://www.xen.org/},
The Xen hypervisor.

363 364
@item [11]
@url{http://www.linux-kvm.org/},
365 366
Kernel Based Virtual Machine (KVM).

367
@item [12]
368 369 370
@url{http://www.greensocs.com/projects/QEMUSystemC},
QEMU-SystemC, a hardware co-simulator.

bellard's avatar
bellard committed
371
@end table