1 |
<html> |
2 |
<head><title>GXemul documentation: Misc.</title> |
3 |
</head> |
4 |
<body bgcolor="#f8f8f8" text="#000000" link="#4040f0" vlink="#404040" alink="#ff0000"> |
5 |
<table border=0 width=100% bgcolor="#d0d0d0"><tr> |
6 |
<td width=100% align=center valign=center><table border=0 width=100%><tr> |
7 |
<td align="left" valign=center bgcolor="#d0efff"><font color="#6060e0" size="6"> |
8 |
<b>GXemul documentation:</b></font> |
9 |
<font color="#000000" size="6"><b>Misc.</b> |
10 |
</font></td></tr></table></td></tr></table><p> |
11 |
<!-- The first 10 lines are cut away by the homepage updating script. --> |
12 |
|
13 |
|
14 |
<!-- |
15 |
|
16 |
$Id: misc.html,v 1.40 2005/04/16 00:29:45 debug Exp $ |
17 |
|
18 |
Copyright (C) 2003-2005 Anders Gavare. All rights reserved. |
19 |
|
20 |
Redistribution and use in source and binary forms, with or without |
21 |
modification, are permitted provided that the following conditions are met: |
22 |
|
23 |
1. Redistributions of source code must retain the above copyright |
24 |
notice, this list of conditions and the following disclaimer. |
25 |
2. Redistributions in binary form must reproduce the above copyright |
26 |
notice, this list of conditions and the following disclaimer in the |
27 |
documentation and/or other materials provided with the distribution. |
28 |
3. The name of the author may not be used to endorse or promote products |
29 |
derived from this software without specific prior written permission. |
30 |
|
31 |
THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND |
32 |
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
33 |
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
34 |
ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE |
35 |
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
36 |
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
37 |
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
38 |
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
39 |
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
40 |
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
41 |
SUCH DAMAGE. |
42 |
|
43 |
--> |
44 |
|
45 |
<a href="./">Back to the index</a> |
46 |
|
47 |
<p><br> |
48 |
<h2>Misc.</h2> |
49 |
|
50 |
<p> |
51 |
<ul> |
52 |
<li><a href="#networking">Networking</a> |
53 |
<li><a href="#portmips">Porting operating systems to MIPS using GXemul</a> |
54 |
<li><a href="#compilercontruct">Using GXemul in compiler contruction courses</a> |
55 |
<li><a href="#disk">How to start the emulator with a disk image</a> |
56 |
<li><a href="#largeimages">How to extract large gzipped disk images</a> |
57 |
<li><a href="#userland">Running userland binaries</a> |
58 |
<li><a href="#promdump">Using a PROM dump from a real machine</a> |
59 |
</ul> |
60 |
|
61 |
|
62 |
|
63 |
|
64 |
|
65 |
|
66 |
|
67 |
<p><br> |
68 |
<a name="networking"></a> |
69 |
<h3>Networking:</h3> |
70 |
|
71 |
It is possible to let the guest OS running inside the emulator get access to |
72 |
the Internet. If you are interested in the technical details, and the |
73 |
reasons why networking is implemented in the emulator the way it currently |
74 |
is implemented, you might want to read the <a href="technical.html#net"> |
75 |
networking section in the technical documentation</a>. |
76 |
<p> |
77 |
The guest OS running inside the emulator uses a private IPv4 address, such |
78 |
as 10.0.0.1, and the emulator acts as a NAT-like gateway/firewall at IPv4 |
79 |
address 10.0.0.254. To the outside world it will seem like it is the host's |
80 |
OS that connects to other machines on the internet, not the guest OS. |
81 |
<p> |
82 |
<font color="#ff0000">NOTE: This is still experimental! |
83 |
As of 2004-07-21, ARP + ICMP + UDP + TCP are emulated well enough to let |
84 |
NetBSD and OpenBSD install via ftp, and use the network for many normal |
85 |
activities, but not everything works yet.</font> |
86 |
|
87 |
|
88 |
|
89 |
|
90 |
|
91 |
|
92 |
<p><br> |
93 |
<a name="portmips"></a> |
94 |
<h3>Porting operating systems to MIPS using GXemul:</h3> |
95 |
|
96 |
Is this a good idea? The answer is yes and no, depending on what you are |
97 |
trying to port to. If you are developing an operating system or operating |
98 |
system kernel of your own, and wish to target MIPS-like systems in general, |
99 |
then the answer might be yes, for experimental purposes. |
100 |
|
101 |
<p> |
102 |
However, if you think that you can port an operating system |
103 |
to, say, the Silicon Graphics machine mode of GXemul and hope that your |
104 |
operating system will run on a real SGI machine, then you will most |
105 |
likely fail. GXemul simply does not emulate things well enough for that to work. |
106 |
Another example would be specific CPU details; if your code depends on, |
107 |
say, R10000 specifics, chances are that GXemul will not be sufficient. |
108 |
|
109 |
<p> |
110 |
In many cases, hardware devices in GXemul are only implemented well |
111 |
enough to fool eg. NetBSD that they are working correctly, while in |
112 |
fact they don't work very much at all. Please keep this in mind, if you plan |
113 |
to use GXemul when porting your code to MIPS. |
114 |
|
115 |
|
116 |
|
117 |
|
118 |
|
119 |
|
120 |
|
121 |
<p><br> |
122 |
<a name="compilercontruct"></a> |
123 |
<h3>Using GXemul in compiler contruction courses:</h3> |
124 |
|
125 |
If you are learning how to write a compiler, and wish to target a |
126 |
realistic target platform, then MIPS (as emulated by GXemul) |
127 |
might be a suitable choice. |
128 |
|
129 |
<ul> |
130 |
<li><b>(+)</b> Your compiler needs to output real assembly |
131 |
language code, which the assembler (eg gas, the GNU assembler) can |
132 |
then compile into object format, and then you need to link this |
133 |
into an executable image. This is much closer to how things work |
134 |
in real life than running assembly language listings in a simulator |
135 |
(eg SPIM). |
136 |
<p> |
137 |
<li><b>(-)</b> GXemul does not simulate out-of-order |
138 |
execution, penalties related to instruction scheduling, or |
139 |
load-delays, so it cannot be used to create optimizing compilers |
140 |
that take advantage of such processor features. GXemul keeps |
141 |
track of the number of instructions executed, but that's it. |
142 |
</ul> |
143 |
|
144 |
|
145 |
|
146 |
|
147 |
|
148 |
|
149 |
<p><br> |
150 |
<a name="disk"></a> |
151 |
<h3>How to start the emulator with a disk image:</h3> |
152 |
|
153 |
Add <i>-d [prefixes:]diskimagefilename</i> to the command line, where prefixes |
154 |
are one or more single-character options. Run <b>gxemul -h</b> |
155 |
to get a list of possible options. |
156 |
|
157 |
<p> |
158 |
Here are some examples. If you want to run a NetBSD/pmax kernel on an |
159 |
emulated DECstation machine, you would use a command line such as this: |
160 |
<pre> |
161 |
$ <b>gxemul -E dec -e 3max -b -d pmax_diskimage.fs netbsd-pmax-INSTALL</b> |
162 |
</pre> |
163 |
<p> |
164 |
NOTE: For some emulation modes, such as the DECstation mode, you do |
165 |
<i>not</i> have to specify the name of the kernel, if the disk image is |
166 |
bootable! |
167 |
<p> |
168 |
It is possible to have more than one disk. For each -d argument, a disk |
169 |
image is added; the first will be SCSI target 0, the second will be target 1, and so on, |
170 |
unless you specify explicitly which ID number the devices should have. |
171 |
<pre> |
172 |
$ <b>gxemul -E dec -e 3max -b -d disk0.raw -d disk1.raw -d 5:disk2.raw netbsd-pmax-INSTALL</b> |
173 |
</pre> |
174 |
Note: In the example above, disk2.raw will get scsi id 5. |
175 |
<p> |
176 |
If a filename has a 'c' prefix, or ends with ".iso", then it is assumed to be |
177 |
a CDROM device (this can be overridden with a 'd' prefix, to force a read/write disk). |
178 |
For example, the following command would start the emulator with two |
179 |
CDROM images, and one harddisk image: |
180 |
<pre> |
181 |
$ <b>gxemul -E dec -e 3max -b -d image.iso -d disk0.img -d c:second_cdrom.img netbsd-pmax-INSTALL</b> |
182 |
</pre> |
183 |
Usually, the device with the lowest id becomes the boot device. To override |
184 |
this, add a 'b' prefix to one of the devices: |
185 |
<pre> |
186 |
$ <b>gxemul -E dec -e 3max -b -d rootdisk.img -d bc:install-cd.iso name_of_kernel</b> |
187 |
</pre> |
188 |
If you have a physical CD-ROM drive on the host machine, say /dev/cd0c, you can |
189 |
use it as a CD-ROM directly accessible from within the emulator: |
190 |
<pre> |
191 |
$ <b>gxemul -E dec -e 3max -b -d rootdisk.img -d bc:/dev/cd0c name_of_kernel</b> |
192 |
</pre> |
193 |
It is probably possible to use harddisks as well this way, but I would not |
194 |
recommend it. |
195 |
<p> |
196 |
Using emulated tape drives is a bit more complicated than disks, because a |
197 |
tape can be made up of several "files" with space in between. The solution |
198 |
I have choosen is to have one file in the host's file system space for each |
199 |
tape file. The prefix for using tapes is 't', and the filename given is |
200 |
for the <i>first</i> file on that tape (number zero, implicitly). For |
201 |
files following file nr 0, a dot and the filenumber is appended to the |
202 |
filename. |
203 |
<p> |
204 |
As an example, starting the emulator with |
205 |
<pre> |
206 |
<b>-d t4:mytape.img</b> |
207 |
</pre> |
208 |
will cause SCSI id 4 to be a tape device, using the following file number |
209 |
to name translation scheme: |
210 |
<p> |
211 |
<center> |
212 |
<table border="0"> |
213 |
<tr> |
214 |
<td><b>File number:</b></td> |
215 |
<td><b>File name in the host's filesystem:</b></td> |
216 |
</tr> |
217 |
<tr> |
218 |
<td align="center">0</td> |
219 |
<td align="left">mytape.img</td> |
220 |
</tr> |
221 |
<tr> |
222 |
<td align="center">1</td> |
223 |
<td align="left">mytape.img.1</td> |
224 |
</tr> |
225 |
<tr> |
226 |
<td align="center">2</td> |
227 |
<td align="left">mytape.img.2</td> |
228 |
</tr> |
229 |
<tr> |
230 |
<td align="center">..</td> |
231 |
<td align="left">..</td> |
232 |
</tr> |
233 |
</table> |
234 |
</center> |
235 |
<p> |
236 |
If you already have a number of tape files, which should be placed on the |
237 |
same emulated tape, then you might not want to rename all those files. |
238 |
Use symbolic links instead (ln -s). |
239 |
<p> |
240 |
There is another advantage to using symbolic links for tape filenames: |
241 |
every time a tape is rewound, it is reopened using the filename given |
242 |
on the command line. By changing what the symbolic name points to, |
243 |
you can "switch tapes" without quiting and restarting the emulator. |
244 |
|
245 |
|
246 |
|
247 |
<p><br> |
248 |
<a name="largeimages"></a> |
249 |
<h3>How to extract large gzipped disk images:</h3> |
250 |
|
251 |
Unix filesystems usually support large files with "holes". Holes are |
252 |
zero-filled blocks that don't actually exist on disk. This is very |
253 |
practical for emulated disk images, as it is possible to create a very |
254 |
large disk image without using up much space at all. |
255 |
|
256 |
<p> |
257 |
Using gzip and gunzip on disk images can be <i>very</i> slow, as these |
258 |
files can be multiple gigabytes large, but this is usually necessary for |
259 |
transfering disk images over the internet. If you receive a gzipped disk |
260 |
image, say disk.img.gz, and run a naive |
261 |
<p> |
262 |
<pre> |
263 |
$ <b>gunzip disk.img.gz</b> |
264 |
</pre> |
265 |
<p> |
266 |
on it, you will not end up with an optimized file unless |
267 |
gunzip supports that. (In my experiments, it doesn't.) In plain English, |
268 |
if you type <b>ls -l</b> and the filesize is 9 GB, it will actually occupy |
269 |
9 GB of disk space! This is often unacceptable. |
270 |
<p> |
271 |
Using a simple tool which only writes blocks that are non-zero, a lot of |
272 |
space can be saved. Compile the program cp_removeblocks in the |
273 |
experiments/ directory, and type: |
274 |
<p> |
275 |
<pre> |
276 |
$ <b>gunzip -c disk.img.gz | cp_removeblocks /dev/stdin disk.img</b> |
277 |
</pre> |
278 |
|
279 |
<p> |
280 |
This will give you a disk.img which looks like it is 9 GB, and works like |
281 |
the real file, but the holes are not written out to the disk. (You can see |
282 |
this by running for example <b>du disk.img</b> to see the physical block |
283 |
count.) |
284 |
|
285 |
|
286 |
|
287 |
<p><br> |
288 |
<a name="userland"></a> |
289 |
<h3>Running userland binaries:</h3> |
290 |
|
291 |
You can run (some) userland programs as well. This will not emulate any |
292 |
particular machine, but instead try to translate syscalls from for example |
293 |
NetBSD/pmax into the host's OS' syscalls. Right now, this is just a |
294 |
proof-of-concept, to show that it would work; there's lots of work left to |
295 |
do to make it actually run useful programs (for example dynamically linked |
296 |
programs). |
297 |
|
298 |
<p> |
299 |
|
300 |
<ul> |
301 |
<li><b>NetBSD/pmax:</b> |
302 |
<br> |
303 |
Running /bin/hostname or /bin/date and similarly trivial |
304 |
programs from the NetBSD/pmax distribution works:<pre> |
305 |
$ <b>gxemul -q -u netbsd/pmax pmax_bin_hostname</b> |
306 |
tab.csbnet.se |
307 |
$ <b>gxemul -q -u netbsd/pmax pmax_bin_date</b> |
308 |
Sun Jan 25 02:26:14 GMT 2004 |
309 |
$ <b>gxemul -q -u netbsd/pmax pmax_bin_sleep</b> |
310 |
usage: pmax_bin_sleep seconds |
311 |
$ <b>gxemul -q -u netbsd/pmax pmax_bin_sleep 5</b> |
312 |
$ <b>gxemul -q -u netbsd/pmax pmax_bin_sync</b> |
313 |
</pre> |
314 |
|
315 |
<p> |
316 |
<li><b>Ultrix:</b> |
317 |
<br> |
318 |
At least /bin/date and /bin/hostname work:<pre> |
319 |
$ <b>gxemul -q -u ultrix ultrix4_bin_date</b> |
320 |
UNIMPLEMENTED ultrix syscall 54 |
321 |
UNIMPLEMENTED ultrix syscall 62 |
322 |
Mon Feb 9 12:50:59 WET 2004 |
323 |
$ <b>gxemul -q -u ultrix ultrix4_bin_hostname</b> |
324 |
tab.csbnet.se |
325 |
</pre> |
326 |
|
327 |
<p> |
328 |
<li><b>NetBSD/powerpc:</b> |
329 |
<br> |
330 |
/bin/sync from NetBSD/macppc works, but probably not much else.<pre> |
331 |
$ <b>gxemul -v -u netbsd/powerpc netbsd-1.6.2-macppc-bin-sync</b> |
332 |
... |
333 |
[ sync() ] |
334 |
[ exit(0) ] |
335 |
cpu_run_deinit(): All CPUs halted. |
336 |
|
337 |
</pre> |
338 |
|
339 |
<p> |
340 |
<li><b>Linux/PPC64:</b> |
341 |
<br> |
342 |
The <a href="http://www-106.ibm.com/developerworks/library/l-ppc/#h13">64-bit Hello World assembly language example</a> |
343 |
on IBM's developerWorks pages runs:<pre> |
344 |
$ <b>ppc64-unknown-linux-as hello-ppc64.s -o hello-ppc64.o</b> |
345 |
$ <b>ppc64-unknown-linux-ld hello-ppc64.o -o hello-ppc64</b> |
346 |
$ <b>gxemul -q -u linux/ppc64 hello-ppc64</b> |
347 |
Hello, world! |
348 |
|
349 |
</pre> |
350 |
|
351 |
</ul> |
352 |
|
353 |
|
354 |
|
355 |
|
356 |
|
357 |
<p><br> |
358 |
<a name="promdump"></a> |
359 |
<h3>Using a PROM dump from a real machine:</h3> |
360 |
|
361 |
Raw PROM images from real machines can, in a few cases, be used in |
362 |
the emulator. ROM code is usually much more sensitive to correctness |
363 |
of the emulator than operating system kernels or userland programs |
364 |
are, so don't expect any PROM image to just magically work. |
365 |
|
366 |
|
367 |
<p> |
368 |
<h4>Dumping the PROM on a DECstation 5000/125:</h4> |
369 |
The image first needs to be extracted from the machine. There are |
370 |
several ways to do this. |
371 |
<ul> |
372 |
<li>Use hardware to read the PROM chip(s) directly. Not easy if you |
373 |
don't have such a hardware reader. |
374 |
<li>Copy the PROM memory range into a file, from a running |
375 |
operating system. You need a running OS, and it must |
376 |
have access to the PROM memory range. NetBSD, for example, |
377 |
doesn't allow that from userland. |
378 |
<li>Hook up a serial console and dump using the PROM's own dump |
379 |
command. |
380 |
</ul> |
381 |
<p> |
382 |
The easiest way is to hook up a serial console. The terminal must be |
383 |
able to capture output to a file. |
384 |
<p> |
385 |
These are approximately the commands that I used: |
386 |
<pre> |
387 |
>><b>cnfg</b> <i>Show machine configuration</i> |
388 |
|
389 |
>><b>printenv</b> <i>Show environment variables</i> |
390 |
|
391 |
>><b>setenv more 0</b> <i>This turns off the More messages</i> |
392 |
|
393 |
>><b>e -x 0xbfc00000:0xbfffffff</b> <i>Dump the PROM data</i> |
394 |
</pre> |
395 |
<p> |
396 |
Remember that DECstations are little endian, so if the dump data |
397 |
looks like this: |
398 |
<pre> |
399 |
bfc00000: 0x0bf0007e |
400 |
</pre> |
401 |
then the bytes in memory are actually 0x7e, 0x00, 0xf0, and 0x0b. |
402 |
<p> |
403 |
At 9600 bps, about 10KB can be dumped per minute, so it takes a while. |
404 |
Once enough of the PROM has been dumped, you can press CTRL-C to break out. |
405 |
Then, restore the more environment variable: |
406 |
<pre> |
407 |
>><b>setenv more 24</b> |
408 |
</pre> |
409 |
<p> |
410 |
Now, convert the data you just saved (little-endian words -> bytes), |
411 |
and store in a file. Let's call this file DECstation5000_125_promdump.bin. |
412 |
<pre> |
413 |
$ <b>decprom_dump_txt_to_bin DECstation5000_125_promdump.txt DECstation5000_125_promdump.bin</b> |
414 |
</pre> |
415 |
This binary image can now be used in the emulator: |
416 |
<pre> |
417 |
$ <b>gxemul -E dec -e 3min -Q -M128 -q 0xbfc00000:DECstation5000_125_promdump.bin</b> |
418 |
|
419 |
KN02-BA V5.7e |
420 |
?TFL: 3/scc/access (1:Ln1 reg-12: actual=0x00 xpctd=0x01) [KN02-BA] |
421 |
?TFL: 3/scc/io (1:Ln0 tx bfr not empty. status=0X 0) [KN02-BA] |
422 |
... |
423 |
--More--?TFL: 3/scsi/cntl (CUX, cause= 1000002C) |
424 |
>><b>?</b> |
425 |
? [cmd] |
426 |
boot [[-z #] [-n] #/path [ARG...]] |
427 |
cat SCRPT |
428 |
cnfg [#] |
429 |
d [-bhw] [-S #] RNG VAL |
430 |
e [-bhwcdoux] [-S #] RNG |
431 |
erl [-c] |
432 |
go [ADR] |
433 |
init [#] [-m] [ARG...] |
434 |
ls [#] |
435 |
passwd [-c] [-s] |
436 |
printenv [EVN] |
437 |
restart |
438 |
script SCRPT |
439 |
setenv EVN STR |
440 |
sh [-belvS] [SCRPT] [ARG..] |
441 |
t [-l] #/STR [ARG..] |
442 |
unsetenv EVN |
443 |
>><b>cnfg</b> |
444 |
3: KN02-BA DEC V5.7e TCF0 (128 MB) |
445 |
(enet: 00-00-00-00-00-00) |
446 |
(SCSI = 7) |
447 |
0: PMAG-BA DEC V5.3a TCF0 |
448 |
>><b>printenv</b> |
449 |
boot= |
450 |
testaction=q |
451 |
haltaction=h |
452 |
more=24 |
453 |
#=3 |
454 |
console=* |
455 |
osconsole=3 |
456 |
>> |
457 |
</pre> |
458 |
<i>(Note: at the moment, this doesn't work. I must have broken something when |
459 |
fixing something else, but this is what it looked like at the time.)</i> |
460 |
<p> |
461 |
During bootup, the PROM complains <i>a lot</i> about hardware failures. |
462 |
That's because the emulator doesn't emulate the hardware well enough yet. |
463 |
<p> |
464 |
The command line options used are: -E dec for DECstation, -e 3min for |
465 |
"model 3" (5000/1xx), -Q to supress the emulator's own PROM |
466 |
call emulation, -M128 for 128MB RAM (because GXemul doesn't correctly |
467 |
emulate memory detection well enough for the PROM to accept, so it will |
468 |
always believe there is 128MB ram anyway), and -q to supress debug messages. |
469 |
The 0xbfc00000 in front of the filename tells GXemul that it is a raw |
470 |
binary file which should be loaded at a specific virtual address. |
471 |
|
472 |
|
473 |
<p><br> |
474 |
<h4>Dumping the PROM on a SGI O2:</h4> |
475 |
|
476 |
The general ideas in this section applies to using ROM images from other |
477 |
machines as well. Besides DECstation, I've also tried this on an SGI IP32 |
478 |
("O2"). |
479 |
<p> |
480 |
For the O2, a suitable command to dump the prom memory range is |
481 |
<pre> |
482 |
>> <b>dump -b 0xBFC00000:0xBFC80000</b> |
483 |
</pre> |
484 |
Make sure you capture all the output (via serial console) into a file, |
485 |
and then run experiments/sgiprom_to_bin on the captured file. |
486 |
<p> |
487 |
(2005-01-16: The emulator doesn't really emulate the IP32 well enough to |
488 |
actually run the PROM image without using special hacks, but it might do |
489 |
so some time in the future.) |
490 |
|
491 |
|
492 |
|
493 |
|
494 |
|
495 |
</p> |
496 |
|
497 |
</body> |
498 |
</html> |