root / qemu-doc.texi @ 97eb5b14
History | View | Annotate | Download (33.5 kB)
1 |
\input texinfo @c -*- texinfo -*- |
---|---|
2 |
|
3 |
@iftex |
4 |
@settitle QEMU CPU Emulator Reference Documentation |
5 |
@titlepage |
6 |
@sp 7 |
7 |
@center @titlefont{QEMU CPU Emulator Reference Documentation} |
8 |
@sp 3 |
9 |
@end titlepage |
10 |
@end iftex |
11 |
|
12 |
@chapter Introduction |
13 |
|
14 |
@section Features |
15 |
|
16 |
QEMU is a FAST! processor emulator. By using dynamic translation it |
17 |
achieves a reasonnable speed while being easy to port on new host |
18 |
CPUs. |
19 |
|
20 |
QEMU has two operating modes: |
21 |
|
22 |
@itemize @minus |
23 |
|
24 |
@item |
25 |
User mode emulation. In this mode, QEMU can launch Linux processes |
26 |
compiled for one CPU on another CPU. Linux system calls are converted |
27 |
because of endianness and 32/64 bit mismatches. The Wine Windows API |
28 |
emulator (@url{http://www.winehq.org}) and the DOSEMU DOS emulator |
29 |
(@url{http://www.dosemu.org}) are the main targets for QEMU. |
30 |
|
31 |
@item |
32 |
Full system emulation. In this mode, QEMU emulates a full |
33 |
system, including a processor and various peripherials. Currently, it |
34 |
is only used to launch an x86 Linux kernel on an x86 Linux system. It |
35 |
enables easier testing and debugging of system code. It can also be |
36 |
used to provide virtual hosting of several virtual PCs on a single |
37 |
server. |
38 |
|
39 |
@end itemize |
40 |
|
41 |
As QEMU requires no host kernel patches to run, it is very safe and |
42 |
easy to use. |
43 |
|
44 |
QEMU generic features: |
45 |
|
46 |
@itemize |
47 |
|
48 |
@item User space only or full system emulation. |
49 |
|
50 |
@item Using dynamic translation to native code for reasonnable speed. |
51 |
|
52 |
@item Working on x86 and PowerPC hosts. Being tested on ARM, Sparc32, Alpha and S390. |
53 |
|
54 |
@item Self-modifying code support. |
55 |
|
56 |
@item Precise exceptions support. |
57 |
|
58 |
@item The virtual CPU is a library (@code{libqemu}) which can be used |
59 |
in other projects. |
60 |
|
61 |
@end itemize |
62 |
|
63 |
QEMU user mode emulation features: |
64 |
@itemize |
65 |
@item Generic Linux system call converter, including most ioctls. |
66 |
|
67 |
@item clone() emulation using native CPU clone() to use Linux scheduler for threads. |
68 |
|
69 |
@item Accurate signal handling by remapping host signals to target signals. |
70 |
@end itemize |
71 |
@end itemize |
72 |
|
73 |
QEMU full system emulation features: |
74 |
@itemize |
75 |
@item QEMU can either use a full software MMU for maximum portability or use the host system call mmap() to simulate the target MMU. |
76 |
@end itemize |
77 |
|
78 |
@section x86 emulation |
79 |
|
80 |
QEMU x86 target features: |
81 |
|
82 |
@itemize |
83 |
|
84 |
@item The virtual x86 CPU supports 16 bit and 32 bit addressing with segmentation. |
85 |
LDT/GDT and IDT are emulated. VM86 mode is also supported to run DOSEMU. |
86 |
|
87 |
@item Support of host page sizes bigger than 4KB in user mode emulation. |
88 |
|
89 |
@item QEMU can emulate itself on x86. |
90 |
|
91 |
@item An extensive Linux x86 CPU test program is included @file{tests/test-i386}. |
92 |
It can be used to test other x86 virtual CPUs. |
93 |
|
94 |
@end itemize |
95 |
|
96 |
Current QEMU limitations: |
97 |
|
98 |
@itemize |
99 |
|
100 |
@item No SSE/MMX support (yet). |
101 |
|
102 |
@item No x86-64 support. |
103 |
|
104 |
@item IPC syscalls are missing. |
105 |
|
106 |
@item The x86 segment limits and access rights are not tested at every |
107 |
memory access. |
108 |
|
109 |
@item On non x86 host CPUs, @code{double}s are used instead of the non standard |
110 |
10 byte @code{long double}s of x86 for floating point emulation to get |
111 |
maximum performances. |
112 |
|
113 |
@item Some priviledged instructions or behaviors are missing, especially for segment protection testing (yet). |
114 |
|
115 |
@end itemize |
116 |
|
117 |
@section ARM emulation |
118 |
|
119 |
@itemize |
120 |
|
121 |
@item ARM emulation can currently launch small programs while using the |
122 |
generic dynamic code generation architecture of QEMU. |
123 |
|
124 |
@item No FPU support (yet). |
125 |
|
126 |
@item No automatic regression testing (yet). |
127 |
|
128 |
@end itemize |
129 |
|
130 |
@section SPARC emulation |
131 |
|
132 |
The SPARC emulation is currently in development. |
133 |
|
134 |
@chapter Installation |
135 |
|
136 |
If you want to compile QEMU, please read the @file{README} which gives |
137 |
the related information. Otherwise just download the binary |
138 |
distribution (@file{qemu-XXX-i386.tar.gz}) and untar it as root in |
139 |
@file{/}: |
140 |
|
141 |
@example |
142 |
su |
143 |
cd / |
144 |
tar zxvf /tmp/qemu-XXX-i386.tar.gz |
145 |
@end example |
146 |
|
147 |
@chapter QEMU User space emulator invocation |
148 |
|
149 |
@section Quick Start |
150 |
|
151 |
In order to launch a Linux process, QEMU needs the process executable |
152 |
itself and all the target (x86) dynamic libraries used by it. |
153 |
|
154 |
@itemize |
155 |
|
156 |
@item On x86, you can just try to launch any process by using the native |
157 |
libraries: |
158 |
|
159 |
@example |
160 |
qemu-i386 -L / /bin/ls |
161 |
@end example |
162 |
|
163 |
@code{-L /} tells that the x86 dynamic linker must be searched with a |
164 |
@file{/} prefix. |
165 |
|
166 |
@item Since QEMU is also a linux process, you can launch qemu with qemu (NOTE: you can only do that if you compiled QEMU from the sources): |
167 |
|
168 |
@example |
169 |
qemu-i386 -L / qemu-i386 -L / /bin/ls |
170 |
@end example |
171 |
|
172 |
@item On non x86 CPUs, you need first to download at least an x86 glibc |
173 |
(@file{qemu-runtime-i386-XXX-.tar.gz} on the QEMU web page). Ensure that |
174 |
@code{LD_LIBRARY_PATH} is not set: |
175 |
|
176 |
@example |
177 |
unset LD_LIBRARY_PATH |
178 |
@end example |
179 |
|
180 |
Then you can launch the precompiled @file{ls} x86 executable: |
181 |
|
182 |
@example |
183 |
qemu-i386 tests/i386/ls |
184 |
@end example |
185 |
You can look at @file{qemu-binfmt-conf.sh} so that |
186 |
QEMU is automatically launched by the Linux kernel when you try to |
187 |
launch x86 executables. It requires the @code{binfmt_misc} module in the |
188 |
Linux kernel. |
189 |
|
190 |
@item The x86 version of QEMU is also included. You can try weird things such as: |
191 |
@example |
192 |
qemu-i386 /usr/local/qemu-i386/bin/qemu-i386 /usr/local/qemu-i386/bin/ls-i386 |
193 |
@end example |
194 |
|
195 |
@end itemize |
196 |
|
197 |
@section Wine launch |
198 |
|
199 |
@itemize |
200 |
|
201 |
@item Ensure that you have a working QEMU with the x86 glibc |
202 |
distribution (see previous section). In order to verify it, you must be |
203 |
able to do: |
204 |
|
205 |
@example |
206 |
qemu-i386 /usr/local/qemu-i386/bin/ls-i386 |
207 |
@end example |
208 |
|
209 |
@item Download the binary x86 Wine install |
210 |
(@file{qemu-XXX-i386-wine.tar.gz} on the QEMU web page). |
211 |
|
212 |
@item Configure Wine on your account. Look at the provided script |
213 |
@file{/usr/local/qemu-i386/bin/wine-conf.sh}. Your previous |
214 |
@code{$@{HOME@}/.wine} directory is saved to @code{$@{HOME@}/.wine.org}. |
215 |
|
216 |
@item Then you can try the example @file{putty.exe}: |
217 |
|
218 |
@example |
219 |
qemu-i386 /usr/local/qemu-i386/wine/bin/wine /usr/local/qemu-i386/wine/c/Program\ Files/putty.exe |
220 |
@end example |
221 |
|
222 |
@end itemize |
223 |
|
224 |
@section Command line options |
225 |
|
226 |
@example |
227 |
usage: qemu-i386 [-h] [-d] [-L path] [-s size] program [arguments...] |
228 |
@end example |
229 |
|
230 |
@table @option |
231 |
@item -h |
232 |
Print the help |
233 |
@item -L path |
234 |
Set the x86 elf interpreter prefix (default=/usr/local/qemu-i386) |
235 |
@item -s size |
236 |
Set the x86 stack size in bytes (default=524288) |
237 |
@end table |
238 |
|
239 |
Debug options: |
240 |
|
241 |
@table @option |
242 |
@item -d |
243 |
Activate log (logfile=/tmp/qemu.log) |
244 |
@item -p pagesize |
245 |
Act as if the host page size was 'pagesize' bytes |
246 |
@end table |
247 |
|
248 |
@chapter QEMU System emulator invocation |
249 |
|
250 |
@section Introduction |
251 |
|
252 |
@c man begin DESCRIPTION |
253 |
|
254 |
The QEMU System emulator simulates a complete PC. It can either boot |
255 |
directly a Linux kernel (without any BIOS or boot loader) or boot like a |
256 |
real PC with the included BIOS. |
257 |
|
258 |
In order to meet specific user needs, two versions of QEMU are |
259 |
available: |
260 |
|
261 |
@enumerate |
262 |
|
263 |
@item |
264 |
@code{qemu-fast} uses the host Memory Management Unit (MMU) to simulate |
265 |
the x86 MMU. It is @emph{fast} but has limitations because the whole 4 GB |
266 |
address space cannot be used and some memory mapped peripherials |
267 |
cannot be emulated accurately yet. Therefore, a specific Linux kernel |
268 |
must be used (@xref{linux_compile}). |
269 |
|
270 |
@item |
271 |
@code{qemu} uses a software MMU. It is about @emph{two times |
272 |
slower} but gives a more accurate emulation. |
273 |
|
274 |
@end enumerate |
275 |
|
276 |
QEMU emulates the following PC peripherials: |
277 |
|
278 |
@itemize @minus |
279 |
@item |
280 |
VGA (hardware level, including all non standard modes) |
281 |
@item |
282 |
PS/2 mouse and keyboard |
283 |
@item |
284 |
2 IDE interfaces with hard disk and CD-ROM support |
285 |
@item |
286 |
NE2000 network adapter (port=0x300, irq=9) |
287 |
@item |
288 |
Serial port |
289 |
@item |
290 |
Soundblaster 16 card |
291 |
@item |
292 |
PIC (interrupt controler) |
293 |
@item |
294 |
PIT (timers) |
295 |
@item |
296 |
CMOS memory |
297 |
@end itemize |
298 |
|
299 |
@c man end |
300 |
|
301 |
@section Quick Start |
302 |
|
303 |
Download and uncompress the linux image (@file{linux.img}) and type: |
304 |
|
305 |
@example |
306 |
qemu linux.img |
307 |
@end example |
308 |
|
309 |
Linux should boot and give you a prompt. |
310 |
|
311 |
@section Direct Linux Boot and Network emulation |
312 |
|
313 |
This section explains how to launch a Linux kernel inside QEMU without |
314 |
having to make a full bootable image. It is very useful for fast Linux |
315 |
kernel testing. The QEMU network configuration is also explained. |
316 |
|
317 |
@enumerate |
318 |
@item |
319 |
Download the archive @file{linux-test-xxx.tar.gz} containing a Linux |
320 |
kernel and a disk image. |
321 |
|
322 |
@item Optional: If you want network support (for example to launch X11 examples), you |
323 |
must copy the script @file{qemu-ifup} in @file{/etc} and configure |
324 |
properly @code{sudo} so that the command @code{ifconfig} contained in |
325 |
@file{qemu-ifup} can be executed as root. You must verify that your host |
326 |
kernel supports the TUN/TAP network interfaces: the device |
327 |
@file{/dev/net/tun} must be present. |
328 |
|
329 |
When network is enabled, there is a virtual network connection between |
330 |
the host kernel and the emulated kernel. The emulated kernel is seen |
331 |
from the host kernel at IP address 172.20.0.2 and the host kernel is |
332 |
seen from the emulated kernel at IP address 172.20.0.1. |
333 |
|
334 |
@item Launch @code{qemu.sh}. You should have the following output: |
335 |
|
336 |
@example |
337 |
> ./qemu.sh |
338 |
Connected to host network interface: tun0 |
339 |
Linux version 2.4.21 (bellard@voyager.localdomain) (gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)) #5 Tue Nov 11 18:18:53 CET 2003 |
340 |
BIOS-provided physical RAM map: |
341 |
BIOS-e801: 0000000000000000 - 000000000009f000 (usable) |
342 |
BIOS-e801: 0000000000100000 - 0000000002000000 (usable) |
343 |
32MB LOWMEM available. |
344 |
On node 0 totalpages: 8192 |
345 |
zone(0): 4096 pages. |
346 |
zone(1): 4096 pages. |
347 |
zone(2): 0 pages. |
348 |
Kernel command line: root=/dev/hda sb=0x220,5,1,5 ide2=noprobe ide3=noprobe ide4=noprobe ide5=noprobe console=ttyS0 |
349 |
ide_setup: ide2=noprobe |
350 |
ide_setup: ide3=noprobe |
351 |
ide_setup: ide4=noprobe |
352 |
ide_setup: ide5=noprobe |
353 |
Initializing CPU#0 |
354 |
Detected 2399.621 MHz processor. |
355 |
Console: colour EGA 80x25 |
356 |
Calibrating delay loop... 4744.80 BogoMIPS |
357 |
Memory: 28872k/32768k available (1210k kernel code, 3508k reserved, 266k data, 64k init, 0k highmem) |
358 |
Dentry cache hash table entries: 4096 (order: 3, 32768 bytes) |
359 |
Inode cache hash table entries: 2048 (order: 2, 16384 bytes) |
360 |
Mount cache hash table entries: 512 (order: 0, 4096 bytes) |
361 |
Buffer-cache hash table entries: 1024 (order: 0, 4096 bytes) |
362 |
Page-cache hash table entries: 8192 (order: 3, 32768 bytes) |
363 |
CPU: Intel Pentium Pro stepping 03 |
364 |
Checking 'hlt' instruction... OK. |
365 |
POSIX conformance testing by UNIFIX |
366 |
Linux NET4.0 for Linux 2.4 |
367 |
Based upon Swansea University Computer Society NET3.039 |
368 |
Initializing RT netlink socket |
369 |
apm: BIOS not found. |
370 |
Starting kswapd |
371 |
Journalled Block Device driver loaded |
372 |
Detected PS/2 Mouse Port. |
373 |
pty: 256 Unix98 ptys configured |
374 |
Serial driver version 5.05c (2001-07-08) with no serial options enabled |
375 |
ttyS00 at 0x03f8 (irq = 4) is a 16450 |
376 |
ne.c:v1.10 9/23/94 Donald Becker (becker@scyld.com) |
377 |
Last modified Nov 1, 2000 by Paul Gortmaker |
378 |
NE*000 ethercard probe at 0x300: 52 54 00 12 34 56 |
379 |
eth0: NE2000 found at 0x300, using IRQ 9. |
380 |
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize |
381 |
Uniform Multi-Platform E-IDE driver Revision: 7.00beta4-2.4 |
382 |
ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx |
383 |
hda: QEMU HARDDISK, ATA DISK drive |
384 |
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 |
385 |
hda: attached ide-disk driver. |
386 |
hda: 20480 sectors (10 MB) w/256KiB Cache, CHS=20/16/63 |
387 |
Partition check: |
388 |
hda: |
389 |
Soundblaster audio driver Copyright (C) by Hannu Savolainen 1993-1996 |
390 |
NET4: Linux TCP/IP 1.0 for NET4.0 |
391 |
IP Protocols: ICMP, UDP, TCP, IGMP |
392 |
IP: routing cache hash table of 512 buckets, 4Kbytes |
393 |
TCP: Hash tables configured (established 2048 bind 4096) |
394 |
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0. |
395 |
EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended |
396 |
VFS: Mounted root (ext2 filesystem). |
397 |
Freeing unused kernel memory: 64k freed |
398 |
|
399 |
Linux version 2.4.21 (bellard@voyager.localdomain) (gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)) #5 Tue Nov 11 18:18:53 CET 2003 |
400 |
|
401 |
QEMU Linux test distribution (based on Redhat 9) |
402 |
|
403 |
Type 'exit' to halt the system |
404 |
|
405 |
sh-2.05b# |
406 |
@end example |
407 |
|
408 |
@item |
409 |
Then you can play with the kernel inside the virtual serial console. You |
410 |
can launch @code{ls} for example. Type @key{Ctrl-a h} to have an help |
411 |
about the keys you can type inside the virtual serial console. In |
412 |
particular, use @key{Ctrl-a x} to exit QEMU and use @key{Ctrl-a b} as |
413 |
the Magic SysRq key. |
414 |
|
415 |
@item |
416 |
If the network is enabled, launch the script @file{/etc/linuxrc} in the |
417 |
emulator (don't forget the leading dot): |
418 |
@example |
419 |
. /etc/linuxrc |
420 |
@end example |
421 |
|
422 |
Then enable X11 connections on your PC from the emulated Linux: |
423 |
@example |
424 |
xhost +172.20.0.2 |
425 |
@end example |
426 |
|
427 |
You can now launch @file{xterm} or @file{xlogo} and verify that you have |
428 |
a real Virtual Linux system ! |
429 |
|
430 |
@end enumerate |
431 |
|
432 |
NOTES: |
433 |
@enumerate |
434 |
@item |
435 |
A 2.5.74 kernel is also included in the archive. Just |
436 |
replace the bzImage in qemu.sh to try it. |
437 |
|
438 |
@item |
439 |
qemu creates a temporary file in @var{$QEMU_TMPDIR} (@file{/tmp} is the |
440 |
default) containing all the simulated PC memory. If possible, try to use |
441 |
a temporary directory using the tmpfs filesystem to avoid too many |
442 |
unnecessary disk accesses. |
443 |
|
444 |
@item |
445 |
In order to exit cleanly from qemu, you can do a @emph{shutdown} inside |
446 |
qemu. qemu will automatically exit when the Linux shutdown is done. |
447 |
|
448 |
@item |
449 |
You can boot slightly faster by disabling the probe of non present IDE |
450 |
interfaces. To do so, add the following options on the kernel command |
451 |
line: |
452 |
@example |
453 |
ide1=noprobe ide2=noprobe ide3=noprobe ide4=noprobe ide5=noprobe |
454 |
@end example |
455 |
|
456 |
@item |
457 |
The example disk image is a modified version of the one made by Kevin |
458 |
Lawton for the plex86 Project (@url{www.plex86.org}). |
459 |
|
460 |
@end enumerate |
461 |
|
462 |
@section Invocation |
463 |
|
464 |
@example |
465 |
@c man begin SYNOPSIS |
466 |
usage: qemu [options] [disk_image] |
467 |
@c man end |
468 |
@end example |
469 |
|
470 |
@c man begin OPTIONS |
471 |
@var{disk_image} is a raw hard disk image for IDE hard disk 0. |
472 |
|
473 |
General options: |
474 |
@table @option |
475 |
@item -fda file |
476 |
@item -fdb file |
477 |
Use @var{file} as floppy disk 0/1 image (@xref{disk_images}). |
478 |
|
479 |
@item -hda file |
480 |
@item -hdb file |
481 |
@item -hdc file |
482 |
@item -hdd file |
483 |
Use @var{file} as hard disk 0, 1, 2 or 3 image (@xref{disk_images}). |
484 |
|
485 |
@item -cdrom file |
486 |
Use @var{file} as CD-ROM image (you cannot use @option{-hdc} and and |
487 |
@option{-cdrom} at the same time). |
488 |
|
489 |
@item -boot [a|b|c|d] |
490 |
Boot on floppy (a, b), hard disk (c) or CD-ROM (d). Hard disk boot is |
491 |
the default. |
492 |
|
493 |
@item -snapshot |
494 |
Write to temporary files instead of disk image files. In this case, |
495 |
the raw disk image you use is not written back. You can however force |
496 |
the write back by pressing @key{C-a s} (@xref{disk_images}). |
497 |
|
498 |
@item -m megs |
499 |
Set virtual RAM size to @var{megs} megabytes. |
500 |
|
501 |
@item -n script |
502 |
Set network init script [default=/etc/qemu-ifup]. This script is |
503 |
launched to configure the host network interface (usually tun0) |
504 |
corresponding to the virtual NE2000 card. |
505 |
|
506 |
@item -initrd file |
507 |
Use @var{file} as initial ram disk. |
508 |
|
509 |
@item -tun-fd fd |
510 |
Assumes @var{fd} talks to tap/tun and use it. Read |
511 |
@url{http://bellard.org/qemu/tetrinet.html} to have an example of its |
512 |
use. |
513 |
|
514 |
@item -nographic |
515 |
|
516 |
Normally, QEMU uses SDL to display the VGA output. With this option, |
517 |
you can totally disable graphical output so that QEMU is a simple |
518 |
command line application. The emulated serial port is redirected on |
519 |
the console. Therefore, you can still use QEMU to debug a Linux kernel |
520 |
with a serial console. |
521 |
|
522 |
@end table |
523 |
|
524 |
Linux boot specific (does not require a full PC boot with a BIOS): |
525 |
@table @option |
526 |
|
527 |
@item -kernel bzImage |
528 |
Use @var{bzImage} as kernel image. |
529 |
|
530 |
@item -append cmdline |
531 |
Use @var{cmdline} as kernel command line |
532 |
|
533 |
@item -initrd file |
534 |
Use @var{file} as initial ram disk. |
535 |
|
536 |
@end table |
537 |
|
538 |
Debug options: |
539 |
@table @option |
540 |
@item -s |
541 |
Wait gdb connection to port 1234 (@xref{gdb_usage}). |
542 |
@item -p port |
543 |
Change gdb connection port. |
544 |
@item -d |
545 |
Output log in /tmp/qemu.log |
546 |
@end table |
547 |
|
548 |
During emulation, use @key{C-a h} to get terminal commands: |
549 |
|
550 |
@table @key |
551 |
@item C-a h |
552 |
Print this help |
553 |
@item C-a x |
554 |
Exit emulatior |
555 |
@item C-a s |
556 |
Save disk data back to file (if -snapshot) |
557 |
@item C-a b |
558 |
Send break (magic sysrq) |
559 |
@item C-a C-a |
560 |
Send C-a |
561 |
@end table |
562 |
@c man end |
563 |
|
564 |
@ignore |
565 |
|
566 |
@setfilename qemu |
567 |
@settitle QEMU System Emulator |
568 |
|
569 |
@c man begin SEEALSO |
570 |
The HTML documentation of QEMU for more precise information and Linux |
571 |
user mode emulator invocation. |
572 |
@c man end |
573 |
|
574 |
@c man begin AUTHOR |
575 |
Fabrice Bellard |
576 |
@c man end |
577 |
|
578 |
@end ignore |
579 |
|
580 |
@end ignore |
581 |
@node disk_images |
582 |
@section Disk Images |
583 |
|
584 |
@subsection Raw disk images |
585 |
|
586 |
The disk images can simply be raw images of the hard disk. You can |
587 |
create them with the command: |
588 |
@example |
589 |
dd if=/dev/zero of=myimage bs=1024 count=mysize |
590 |
@end example |
591 |
where @var{myimage} is the image filename and @var{mysize} is its size |
592 |
in kilobytes. |
593 |
|
594 |
@subsection Snapshot mode |
595 |
|
596 |
If you use the option @option{-snapshot}, all disk images are |
597 |
considered as read only. When sectors in written, they are written in |
598 |
a temporary file created in @file{/tmp}. You can however force the |
599 |
write back to the raw disk images by pressing @key{C-a s}. |
600 |
|
601 |
NOTE: The snapshot mode only works with raw disk images. |
602 |
|
603 |
@subsection Copy On Write disk images |
604 |
|
605 |
QEMU also supports user mode Linux |
606 |
(@url{http://user-mode-linux.sourceforge.net/}) Copy On Write (COW) |
607 |
disk images. The COW disk images are much smaller than normal images |
608 |
as they store only modified sectors. They also permit the use of the |
609 |
same disk image template for many users. |
610 |
|
611 |
To create a COW disk images, use the command: |
612 |
|
613 |
@example |
614 |
qemu-mkcow -f myrawimage.bin mycowimage.cow |
615 |
@end example |
616 |
|
617 |
@file{myrawimage.bin} is a raw image you want to use as original disk |
618 |
image. It will never be written to. |
619 |
|
620 |
@file{mycowimage.cow} is the COW disk image which is created by |
621 |
@code{qemu-mkcow}. You can use it directly with the @option{-hdx} |
622 |
options. You must not modify the original raw disk image if you use |
623 |
COW images, as COW images only store the modified sectors from the raw |
624 |
disk image. QEMU stores the original raw disk image name and its |
625 |
modified time in the COW disk image so that chances of mistakes are |
626 |
reduced. |
627 |
|
628 |
If the raw disk image is not read-only, by pressing @key{C-a s} you |
629 |
can flush the COW disk image back into the raw disk image, as in |
630 |
snapshot mode. |
631 |
|
632 |
COW disk images can also be created without a corresponding raw disk |
633 |
image. It is useful to have a big initial virtual disk image without |
634 |
using much disk space. Use: |
635 |
|
636 |
@example |
637 |
qemu-mkcow mycowimage.cow 1024 |
638 |
@end example |
639 |
|
640 |
to create a 1 gigabyte empty COW disk image. |
641 |
|
642 |
NOTES: |
643 |
@enumerate |
644 |
@item |
645 |
COW disk images must be created on file systems supporting |
646 |
@emph{holes} such as ext2 or ext3. |
647 |
@item |
648 |
Since holes are used, the displayed size of the COW disk image is not |
649 |
the real one. To know it, use the @code{ls -ls} command. |
650 |
@end enumerate |
651 |
|
652 |
@node linux_compile |
653 |
@section Linux Kernel Compilation |
654 |
|
655 |
You can use any linux kernel with QEMU. However, if you want to use |
656 |
@code{qemu-fast} to get maximum performances, you should make the |
657 |
following changes to the Linux kernel (only 2.4.x and 2.5.x were |
658 |
tested): |
659 |
|
660 |
@enumerate |
661 |
@item |
662 |
The kernel must be mapped at 0x90000000 (the default is |
663 |
0xc0000000). You must modify only two lines in the kernel source: |
664 |
|
665 |
In @file{include/asm/page.h}, replace |
666 |
@example |
667 |
#define __PAGE_OFFSET (0xc0000000) |
668 |
@end example |
669 |
by |
670 |
@example |
671 |
#define __PAGE_OFFSET (0x90000000) |
672 |
@end example |
673 |
|
674 |
And in @file{arch/i386/vmlinux.lds}, replace |
675 |
@example |
676 |
. = 0xc0000000 + 0x100000; |
677 |
@end example |
678 |
by |
679 |
@example |
680 |
. = 0x90000000 + 0x100000; |
681 |
@end example |
682 |
|
683 |
@item |
684 |
If you want to enable SMP (Symmetric Multi-Processing) support, you |
685 |
must make the following change in @file{include/asm/fixmap.h}. Replace |
686 |
@example |
687 |
#define FIXADDR_TOP (0xffffX000UL) |
688 |
@end example |
689 |
by |
690 |
@example |
691 |
#define FIXADDR_TOP (0xa7ffX000UL) |
692 |
@end example |
693 |
(X is 'e' or 'f' depending on the kernel version). Although you can |
694 |
use an SMP kernel with QEMU, it only supports one CPU. |
695 |
|
696 |
@item |
697 |
If you are not using a 2.5 kernel as host kernel but if you use a target |
698 |
2.5 kernel, you must also ensure that the 'HZ' define is set to 100 |
699 |
(1000 is the default) as QEMU cannot currently emulate timers at |
700 |
frequencies greater than 100 Hz on host Linux systems < 2.5. In |
701 |
@file{include/asm/param.h}, replace: |
702 |
|
703 |
@example |
704 |
# define HZ 1000 /* Internal kernel timer frequency */ |
705 |
@end example |
706 |
by |
707 |
@example |
708 |
# define HZ 100 /* Internal kernel timer frequency */ |
709 |
@end example |
710 |
|
711 |
@end enumerate |
712 |
|
713 |
The file config-2.x.x gives the configuration of the example kernels. |
714 |
|
715 |
Just type |
716 |
@example |
717 |
make bzImage |
718 |
@end example |
719 |
|
720 |
As you would do to make a real kernel. Then you can use with QEMU |
721 |
exactly the same kernel as you would boot on your PC (in |
722 |
@file{arch/i386/boot/bzImage}). |
723 |
|
724 |
@node gdb_usage |
725 |
@section GDB usage |
726 |
|
727 |
QEMU has a primitive support to work with gdb, so that you can do |
728 |
'Ctrl-C' while the virtual machine is running and inspect its state. |
729 |
|
730 |
In order to use gdb, launch qemu with the '-s' option. It will wait for a |
731 |
gdb connection: |
732 |
@example |
733 |
> qemu -s -kernel arch/i386/boot/bzImage -hda root-2.4.20.img -append "root=/dev/hda" |
734 |
Connected to host network interface: tun0 |
735 |
Waiting gdb connection on port 1234 |
736 |
@end example |
737 |
|
738 |
Then launch gdb on the 'vmlinux' executable: |
739 |
@example |
740 |
> gdb vmlinux |
741 |
@end example |
742 |
|
743 |
In gdb, connect to QEMU: |
744 |
@example |
745 |
(gdb) target remote localhost:1234 |
746 |
@end example |
747 |
|
748 |
Then you can use gdb normally. For example, type 'c' to launch the kernel: |
749 |
@example |
750 |
(gdb) c |
751 |
@end example |
752 |
|
753 |
Here are some useful tips in order to use gdb on system code: |
754 |
|
755 |
@enumerate |
756 |
@item |
757 |
Use @code{info reg} to display all the CPU registers. |
758 |
@item |
759 |
Use @code{x/10i $eip} to display the code at the PC position. |
760 |
@item |
761 |
Use @code{set architecture i8086} to dump 16 bit code. Then use |
762 |
@code{x/10i $cs*16+*eip} to dump the code at the PC position. |
763 |
@end enumerate |
764 |
|
765 |
@chapter QEMU Internals |
766 |
|
767 |
@section QEMU compared to other emulators |
768 |
|
769 |
Like bochs [3], QEMU emulates an x86 CPU. But QEMU is much faster than |
770 |
bochs as it uses dynamic compilation and because it uses the host MMU to |
771 |
simulate the x86 MMU. The downside is that currently the emulation is |
772 |
not as accurate as bochs (for example, you cannot currently run Windows |
773 |
inside QEMU). |
774 |
|
775 |
Like Valgrind [2], QEMU does user space emulation and dynamic |
776 |
translation. Valgrind is mainly a memory debugger while QEMU has no |
777 |
support for it (QEMU could be used to detect out of bound memory |
778 |
accesses as Valgrind, but it has no support to track uninitialised data |
779 |
as Valgrind does). The Valgrind dynamic translator generates better code |
780 |
than QEMU (in particular it does register allocation) but it is closely |
781 |
tied to an x86 host and target and has no support for precise exceptions |
782 |
and system emulation. |
783 |
|
784 |
EM86 [4] is the closest project to user space QEMU (and QEMU still uses |
785 |
some of its code, in particular the ELF file loader). EM86 was limited |
786 |
to an alpha host and used a proprietary and slow interpreter (the |
787 |
interpreter part of the FX!32 Digital Win32 code translator [5]). |
788 |
|
789 |
TWIN [6] is a Windows API emulator like Wine. It is less accurate than |
790 |
Wine but includes a protected mode x86 interpreter to launch x86 Windows |
791 |
executables. Such an approach as greater potential because most of the |
792 |
Windows API is executed natively but it is far more difficult to develop |
793 |
because all the data structures and function parameters exchanged |
794 |
between the API and the x86 code must be converted. |
795 |
|
796 |
User mode Linux [7] was the only solution before QEMU to launch a Linux |
797 |
kernel as a process while not needing any host kernel patches. However, |
798 |
user mode Linux requires heavy kernel patches while QEMU accepts |
799 |
unpatched Linux kernels. It would be interesting to compare the |
800 |
performance of the two approaches. |
801 |
|
802 |
The new Plex86 [8] PC virtualizer is done in the same spirit as the QEMU |
803 |
system emulator. It requires a patched Linux kernel to work (you cannot |
804 |
launch the same kernel on your PC), but the patches are really small. As |
805 |
it is a PC virtualizer (no emulation is done except for some priveledged |
806 |
instructions), it has the potential of being faster than QEMU. The |
807 |
downside is that a complicated (and potentially unsafe) host kernel |
808 |
patch is needed. |
809 |
|
810 |
@section Portable dynamic translation |
811 |
|
812 |
QEMU is a dynamic translator. When it first encounters a piece of code, |
813 |
it converts it to the host instruction set. Usually dynamic translators |
814 |
are very complicated and highly CPU dependent. QEMU uses some tricks |
815 |
which make it relatively easily portable and simple while achieving good |
816 |
performances. |
817 |
|
818 |
The basic idea is to split every x86 instruction into fewer simpler |
819 |
instructions. Each simple instruction is implemented by a piece of C |
820 |
code (see @file{op-i386.c}). Then a compile time tool (@file{dyngen}) |
821 |
takes the corresponding object file (@file{op-i386.o}) to generate a |
822 |
dynamic code generator which concatenates the simple instructions to |
823 |
build a function (see @file{op-i386.h:dyngen_code()}). |
824 |
|
825 |
In essence, the process is similar to [1], but more work is done at |
826 |
compile time. |
827 |
|
828 |
A key idea to get optimal performances is that constant parameters can |
829 |
be passed to the simple operations. For that purpose, dummy ELF |
830 |
relocations are generated with gcc for each constant parameter. Then, |
831 |
the tool (@file{dyngen}) can locate the relocations and generate the |
832 |
appriopriate C code to resolve them when building the dynamic code. |
833 |
|
834 |
That way, QEMU is no more difficult to port than a dynamic linker. |
835 |
|
836 |
To go even faster, GCC static register variables are used to keep the |
837 |
state of the virtual CPU. |
838 |
|
839 |
@section Register allocation |
840 |
|
841 |
Since QEMU uses fixed simple instructions, no efficient register |
842 |
allocation can be done. However, because RISC CPUs have a lot of |
843 |
register, most of the virtual CPU state can be put in registers without |
844 |
doing complicated register allocation. |
845 |
|
846 |
@section Condition code optimisations |
847 |
|
848 |
Good CPU condition codes emulation (@code{EFLAGS} register on x86) is a |
849 |
critical point to get good performances. QEMU uses lazy condition code |
850 |
evaluation: instead of computing the condition codes after each x86 |
851 |
instruction, it just stores one operand (called @code{CC_SRC}), the |
852 |
result (called @code{CC_DST}) and the type of operation (called |
853 |
@code{CC_OP}). |
854 |
|
855 |
@code{CC_OP} is almost never explicitely set in the generated code |
856 |
because it is known at translation time. |
857 |
|
858 |
In order to increase performances, a backward pass is performed on the |
859 |
generated simple instructions (see |
860 |
@code{translate-i386.c:optimize_flags()}). When it can be proved that |
861 |
the condition codes are not needed by the next instructions, no |
862 |
condition codes are computed at all. |
863 |
|
864 |
@section CPU state optimisations |
865 |
|
866 |
The x86 CPU has many internal states which change the way it evaluates |
867 |
instructions. In order to achieve a good speed, the translation phase |
868 |
considers that some state information of the virtual x86 CPU cannot |
869 |
change in it. For example, if the SS, DS and ES segments have a zero |
870 |
base, then the translator does not even generate an addition for the |
871 |
segment base. |
872 |
|
873 |
[The FPU stack pointer register is not handled that way yet]. |
874 |
|
875 |
@section Translation cache |
876 |
|
877 |
A 2MByte cache holds the most recently used translations. For |
878 |
simplicity, it is completely flushed when it is full. A translation unit |
879 |
contains just a single basic block (a block of x86 instructions |
880 |
terminated by a jump or by a virtual CPU state change which the |
881 |
translator cannot deduce statically). |
882 |
|
883 |
@section Direct block chaining |
884 |
|
885 |
After each translated basic block is executed, QEMU uses the simulated |
886 |
Program Counter (PC) and other cpu state informations (such as the CS |
887 |
segment base value) to find the next basic block. |
888 |
|
889 |
In order to accelerate the most common cases where the new simulated PC |
890 |
is known, QEMU can patch a basic block so that it jumps directly to the |
891 |
next one. |
892 |
|
893 |
The most portable code uses an indirect jump. An indirect jump makes it |
894 |
easier to make the jump target modification atomic. On some |
895 |
architectures (such as PowerPC), the @code{JUMP} opcode is directly |
896 |
patched so that the block chaining has no overhead. |
897 |
|
898 |
@section Self-modifying code and translated code invalidation |
899 |
|
900 |
Self-modifying code is a special challenge in x86 emulation because no |
901 |
instruction cache invalidation is signaled by the application when code |
902 |
is modified. |
903 |
|
904 |
When translated code is generated for a basic block, the corresponding |
905 |
host page is write protected if it is not already read-only (with the |
906 |
system call @code{mprotect()}). Then, if a write access is done to the |
907 |
page, Linux raises a SEGV signal. QEMU then invalidates all the |
908 |
translated code in the page and enables write accesses to the page. |
909 |
|
910 |
Correct translated code invalidation is done efficiently by maintaining |
911 |
a linked list of every translated block contained in a given page. Other |
912 |
linked lists are also maintained to undo direct block chaining. |
913 |
|
914 |
Although the overhead of doing @code{mprotect()} calls is important, |
915 |
most MSDOS programs can be emulated at reasonnable speed with QEMU and |
916 |
DOSEMU. |
917 |
|
918 |
Note that QEMU also invalidates pages of translated code when it detects |
919 |
that memory mappings are modified with @code{mmap()} or @code{munmap()}. |
920 |
|
921 |
@section Exception support |
922 |
|
923 |
longjmp() is used when an exception such as division by zero is |
924 |
encountered. |
925 |
|
926 |
The host SIGSEGV and SIGBUS signal handlers are used to get invalid |
927 |
memory accesses. The exact CPU state can be retrieved because all the |
928 |
x86 registers are stored in fixed host registers. The simulated program |
929 |
counter is found by retranslating the corresponding basic block and by |
930 |
looking where the host program counter was at the exception point. |
931 |
|
932 |
The virtual CPU cannot retrieve the exact @code{EFLAGS} register because |
933 |
in some cases it is not computed because of condition code |
934 |
optimisations. It is not a big concern because the emulated code can |
935 |
still be restarted in any cases. |
936 |
|
937 |
@section Linux system call translation |
938 |
|
939 |
QEMU includes a generic system call translator for Linux. It means that |
940 |
the parameters of the system calls can be converted to fix the |
941 |
endianness and 32/64 bit issues. The IOCTLs are converted with a generic |
942 |
type description system (see @file{ioctls.h} and @file{thunk.c}). |
943 |
|
944 |
QEMU supports host CPUs which have pages bigger than 4KB. It records all |
945 |
the mappings the process does and try to emulated the @code{mmap()} |
946 |
system calls in cases where the host @code{mmap()} call would fail |
947 |
because of bad page alignment. |
948 |
|
949 |
@section Linux signals |
950 |
|
951 |
Normal and real-time signals are queued along with their information |
952 |
(@code{siginfo_t}) as it is done in the Linux kernel. Then an interrupt |
953 |
request is done to the virtual CPU. When it is interrupted, one queued |
954 |
signal is handled by generating a stack frame in the virtual CPU as the |
955 |
Linux kernel does. The @code{sigreturn()} system call is emulated to return |
956 |
from the virtual signal handler. |
957 |
|
958 |
Some signals (such as SIGALRM) directly come from the host. Other |
959 |
signals are synthetized from the virtual CPU exceptions such as SIGFPE |
960 |
when a division by zero is done (see @code{main.c:cpu_loop()}). |
961 |
|
962 |
The blocked signal mask is still handled by the host Linux kernel so |
963 |
that most signal system calls can be redirected directly to the host |
964 |
Linux kernel. Only the @code{sigaction()} and @code{sigreturn()} system |
965 |
calls need to be fully emulated (see @file{signal.c}). |
966 |
|
967 |
@section clone() system call and threads |
968 |
|
969 |
The Linux clone() system call is usually used to create a thread. QEMU |
970 |
uses the host clone() system call so that real host threads are created |
971 |
for each emulated thread. One virtual CPU instance is created for each |
972 |
thread. |
973 |
|
974 |
The virtual x86 CPU atomic operations are emulated with a global lock so |
975 |
that their semantic is preserved. |
976 |
|
977 |
Note that currently there are still some locking issues in QEMU. In |
978 |
particular, the translated cache flush is not protected yet against |
979 |
reentrancy. |
980 |
|
981 |
@section Self-virtualization |
982 |
|
983 |
QEMU was conceived so that ultimately it can emulate itself. Although |
984 |
it is not very useful, it is an important test to show the power of the |
985 |
emulator. |
986 |
|
987 |
Achieving self-virtualization is not easy because there may be address |
988 |
space conflicts. QEMU solves this problem by being an executable ELF |
989 |
shared object as the ld-linux.so ELF interpreter. That way, it can be |
990 |
relocated at load time. |
991 |
|
992 |
@section MMU emulation |
993 |
|
994 |
For system emulation, QEMU uses the mmap() system call to emulate the |
995 |
target CPU MMU. It works as long the emulated OS does not use an area |
996 |
reserved by the host OS (such as the area above 0xc0000000 on x86 |
997 |
Linux). |
998 |
|
999 |
It is planned to add a slower but more precise MMU emulation |
1000 |
with a software MMU. |
1001 |
|
1002 |
@section Bibliography |
1003 |
|
1004 |
@table @asis |
1005 |
|
1006 |
@item [1] |
1007 |
@url{http://citeseer.nj.nec.com/piumarta98optimizing.html}, Optimizing |
1008 |
direct threaded code by selective inlining (1998) by Ian Piumarta, Fabio |
1009 |
Riccardi. |
1010 |
|
1011 |
@item [2] |
1012 |
@url{http://developer.kde.org/~sewardj/}, Valgrind, an open-source |
1013 |
memory debugger for x86-GNU/Linux, by Julian Seward. |
1014 |
|
1015 |
@item [3] |
1016 |
@url{http://bochs.sourceforge.net/}, the Bochs IA-32 Emulator Project, |
1017 |
by Kevin Lawton et al. |
1018 |
|
1019 |
@item [4] |
1020 |
@url{http://www.cs.rose-hulman.edu/~donaldlf/em86/index.html}, the EM86 |
1021 |
x86 emulator on Alpha-Linux. |
1022 |
|
1023 |
@item [5] |
1024 |
@url{http://www.usenix.org/publications/library/proceedings/usenix-nt97/full_papers/chernoff/chernoff.pdf}, |
1025 |
DIGITAL FX!32: Running 32-Bit x86 Applications on Alpha NT, by Anton |
1026 |
Chernoff and Ray Hookway. |
1027 |
|
1028 |
@item [6] |
1029 |
@url{http://www.willows.com/}, Windows API library emulation from |
1030 |
Willows Software. |
1031 |
|
1032 |
@item [7] |
1033 |
@url{http://user-mode-linux.sourceforge.net/}, |
1034 |
The User-mode Linux Kernel. |
1035 |
|
1036 |
@item [8] |
1037 |
@url{http://www.plex86.org/}, |
1038 |
The new Plex86 project. |
1039 |
|
1040 |
@end table |
1041 |
|
1042 |
@chapter Regression Tests |
1043 |
|
1044 |
In the directory @file{tests/}, various interesting testing programs |
1045 |
are available. There are used for regression testing. |
1046 |
|
1047 |
@section @file{test-i386} |
1048 |
|
1049 |
This program executes most of the 16 bit and 32 bit x86 instructions and |
1050 |
generates a text output. It can be compared with the output obtained with |
1051 |
a real CPU or another emulator. The target @code{make test} runs this |
1052 |
program and a @code{diff} on the generated output. |
1053 |
|
1054 |
The Linux system call @code{modify_ldt()} is used to create x86 selectors |
1055 |
to test some 16 bit addressing and 32 bit with segmentation cases. |
1056 |
|
1057 |
The Linux system call @code{vm86()} is used to test vm86 emulation. |
1058 |
|
1059 |
Various exceptions are raised to test most of the x86 user space |
1060 |
exception reporting. |
1061 |
|
1062 |
@section @file{linux-test} |
1063 |
|
1064 |
This program tests various Linux system calls. It is used to verify |
1065 |
that the system call parameters are correctly converted between target |
1066 |
and host CPUs. |
1067 |
|
1068 |
@section @file{hello-i386} |
1069 |
|
1070 |
Very simple statically linked x86 program, just to test QEMU during a |
1071 |
port to a new host CPU. |
1072 |
|
1073 |
@section @file{hello-arm} |
1074 |
|
1075 |
Very simple statically linked ARM program, just to test QEMU during a |
1076 |
port to a new host CPU. |
1077 |
|
1078 |
@section @file{sha1} |
1079 |
|
1080 |
It is a simple benchmark. Care must be taken to interpret the results |
1081 |
because it mostly tests the ability of the virtual CPU to optimize the |
1082 |
@code{rol} x86 instruction and the condition code computations. |
1083 |
|