1 |
dpavlin |
12 |
<html><head><title>Gavare's eXperimental Emulator: Miscellaneous</title> |
2 |
dpavlin |
8 |
<meta name="robots" content="noarchive,nofollow,noindex"></head> |
3 |
dpavlin |
4 |
<body bgcolor="#f8f8f8" text="#000000" link="#4040f0" vlink="#404040" alink="#ff0000"> |
4 |
|
|
<table border=0 width=100% bgcolor="#d0d0d0"><tr> |
5 |
|
|
<td width=100% align=center valign=center><table border=0 width=100%><tr> |
6 |
|
|
<td align="left" valign=center bgcolor="#d0efff"><font color="#6060e0" size="6"> |
7 |
dpavlin |
22 |
<b>Gavare's eXperimental Emulator:</b></font><br> |
8 |
dpavlin |
12 |
<font color="#000000" size="6"><b>Miscellaneous</b> |
9 |
dpavlin |
4 |
</font></td></tr></table></td></tr></table><p> |
10 |
dpavlin |
2 |
|
11 |
|
|
<!-- |
12 |
|
|
|
13 |
dpavlin |
40 |
$Id: misc.html,v 1.68 2007/04/28 09:18:34 debug Exp $ |
14 |
dpavlin |
2 |
|
15 |
dpavlin |
34 |
Copyright (C) 2003-2007 Anders Gavare. All rights reserved. |
16 |
dpavlin |
2 |
|
17 |
|
|
Redistribution and use in source and binary forms, with or without |
18 |
|
|
modification, are permitted provided that the following conditions are met: |
19 |
|
|
|
20 |
|
|
1. Redistributions of source code must retain the above copyright |
21 |
|
|
notice, this list of conditions and the following disclaimer. |
22 |
|
|
2. Redistributions in binary form must reproduce the above copyright |
23 |
|
|
notice, this list of conditions and the following disclaimer in the |
24 |
|
|
documentation and/or other materials provided with the distribution. |
25 |
|
|
3. The name of the author may not be used to endorse or promote products |
26 |
|
|
derived from this software without specific prior written permission. |
27 |
|
|
|
28 |
|
|
THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND |
29 |
|
|
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
30 |
|
|
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
31 |
|
|
ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE |
32 |
|
|
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
33 |
|
|
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
34 |
|
|
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
35 |
|
|
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
36 |
|
|
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
37 |
|
|
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
38 |
|
|
SUCH DAMAGE. |
39 |
|
|
|
40 |
|
|
--> |
41 |
|
|
|
42 |
dpavlin |
12 |
|
43 |
dpavlin |
2 |
<a href="./">Back to the index</a> |
44 |
|
|
|
45 |
|
|
<p><br> |
46 |
dpavlin |
12 |
<h2>Miscellaneous</h2> |
47 |
dpavlin |
2 |
|
48 |
|
|
<p> |
49 |
|
|
<ul> |
50 |
dpavlin |
12 |
<li><a href="#devel">Writing operating system code, or |
51 |
|
|
developing firmware, using GXemul</a> |
52 |
dpavlin |
2 |
<li><a href="#compilercontruct">Using GXemul in compiler contruction courses</a> |
53 |
|
|
<li><a href="#disk">How to start the emulator with a disk image</a> |
54 |
dpavlin |
38 |
<li><a href="#tape_images">How to start the emulator with tape images</a> |
55 |
|
|
<li><a href="#disk_overlays">How to use disk image overlays</a> |
56 |
dpavlin |
20 |
<li><a href="#filexfer">Transfering files to/from the guest OS</a> |
57 |
dpavlin |
2 |
<li><a href="#largeimages">How to extract large gzipped disk images</a> |
58 |
|
|
<li><a href="#userland">Running userland binaries</a> |
59 |
|
|
<li><a href="#promdump">Using a PROM dump from a real machine</a> |
60 |
|
|
</ul> |
61 |
|
|
|
62 |
|
|
|
63 |
|
|
|
64 |
|
|
|
65 |
|
|
|
66 |
|
|
|
67 |
|
|
|
68 |
|
|
|
69 |
|
|
|
70 |
|
|
<p><br> |
71 |
dpavlin |
12 |
<a name="devel"></a> |
72 |
|
|
<h3>Writing operating system code, or developing firmware, using GXemul:</h3> |
73 |
dpavlin |
2 |
|
74 |
dpavlin |
12 |
Is this a good idea? The answer is yes and no, depending on the level of |
75 |
|
|
detail you need in your simulations. If you are developing an operating |
76 |
dpavlin |
20 |
system or operating system kernel of your own, then the emulator can be a |
77 |
|
|
complement to testing on real hardware. |
78 |
dpavlin |
2 |
|
79 |
dpavlin |
20 |
<p>Important things to keep in mind: |
80 |
dpavlin |
2 |
|
81 |
dpavlin |
12 |
<ul> |
82 |
|
|
<li>Porting code to a specific machine mode, e.g. a Silicon Graphics |
83 |
dpavlin |
20 |
machine, using GXemul, will not "magically" cause the code to |
84 |
dpavlin |
12 |
work on a real machine. Sometimes code works in GXemul which doesn't |
85 |
|
|
work on real hardware, sometimes it's the other way around. |
86 |
dpavlin |
2 |
|
87 |
dpavlin |
12 |
<p> |
88 |
|
|
<li>GXemul contains bugs, and many things are not yet implemented. |
89 |
dpavlin |
2 |
|
90 |
dpavlin |
12 |
<p> |
91 |
dpavlin |
20 |
<li><b>Very important!</b> I have only implemented devices in GXemul |
92 |
|
|
to the degree that NetBSD, OpenBSD, Linux, etc don't complain too much. |
93 |
dpavlin |
12 |
<p> |
94 |
dpavlin |
20 |
If you are developing a driver for a device which is emulated by |
95 |
|
|
GXemul, and your driver does not seem to be working, then the |
96 |
|
|
probability of a bug in GXemul's implementation of the device is |
97 |
|
|
very much higher than that of a bug in your driver. |
98 |
|
|
<p> |
99 |
|
|
The device implementations in GXemul are based on the assumption |
100 |
|
|
that the emulated OS is already developed and bug-free. They are |
101 |
|
|
not primarily intended to be used for development of new device |
102 |
|
|
driver code in operating systems, so if you do that, then be |
103 |
|
|
prepared for bugs and inconsitencies. |
104 |
|
|
<p> |
105 |
dpavlin |
12 |
<li>CPU details in GXemul are usually wrong. If your code depends |
106 |
|
|
on, say, R10000 or MIPS64 specifics, chances are that GXemul will |
107 |
dpavlin |
20 |
not be sufficient. One example is different revisions of ISAs; |
108 |
dpavlin |
32 |
some instructions which should trigger an exception on a |
109 |
|
|
real MIPS processor usually execute anyway in GXemul. Another |
110 |
dpavlin |
20 |
example is if userland code tries to access kernel memory; in some |
111 |
|
|
cases there is protection against this, but not in all cases (to get |
112 |
|
|
higher performance). |
113 |
dpavlin |
12 |
<p> |
114 |
|
|
<li>Caches. There is no cache emulation in GXemul right now. Caches |
115 |
|
|
for R2000/R3000 are faked well enough to run NetBSD, Ultrix, etc |
116 |
|
|
in the DECstation emulation mode, but other than that, cache |
117 |
|
|
operations are treated as nops. |
118 |
|
|
</ul> |
119 |
dpavlin |
2 |
|
120 |
dpavlin |
12 |
<p>The bottom line is that GXemul can be useful as yet another way to test |
121 |
|
|
your code during development, but it should not be fully relied on. |
122 |
dpavlin |
2 |
|
123 |
|
|
|
124 |
dpavlin |
12 |
|
125 |
|
|
|
126 |
|
|
|
127 |
|
|
|
128 |
dpavlin |
2 |
<p><br> |
129 |
|
|
<a name="compilercontruct"></a> |
130 |
|
|
<h3>Using GXemul in compiler contruction courses:</h3> |
131 |
|
|
|
132 |
|
|
If you are learning how to write a compiler, and wish to target a |
133 |
dpavlin |
20 |
realistic target platform, then MIPS or ARM (as emulated by GXemul) |
134 |
|
|
might be suitable choices. |
135 |
dpavlin |
2 |
|
136 |
|
|
<ul> |
137 |
|
|
<li><b>(+)</b> Your compiler needs to output real assembly |
138 |
dpavlin |
24 |
language code, which the assembler (e.g. gas, the GNU assembler) can |
139 |
dpavlin |
2 |
then compile into object format, and then you need to link this |
140 |
|
|
into an executable image. This is much closer to how things work |
141 |
|
|
in real life than running assembly language listings in a simulator |
142 |
dpavlin |
24 |
(e.g. SPIM). |
143 |
dpavlin |
2 |
<p> |
144 |
|
|
<li><b>(-)</b> GXemul does not simulate out-of-order |
145 |
|
|
execution, penalties related to instruction scheduling, or |
146 |
|
|
load-delays, so it cannot be used to create optimizing compilers |
147 |
|
|
that take advantage of such processor features. GXemul keeps |
148 |
|
|
track of the number of instructions executed, but that's it. |
149 |
|
|
</ul> |
150 |
|
|
|
151 |
|
|
|
152 |
|
|
|
153 |
|
|
|
154 |
|
|
|
155 |
|
|
|
156 |
|
|
<p><br> |
157 |
|
|
<a name="disk"></a> |
158 |
|
|
<h3>How to start the emulator with a disk image:</h3> |
159 |
|
|
|
160 |
|
|
Add <i>-d [prefixes:]diskimagefilename</i> to the command line, where prefixes |
161 |
|
|
are one or more single-character options. Run <b>gxemul -h</b> |
162 |
|
|
to get a list of possible options. |
163 |
|
|
|
164 |
|
|
<p> |
165 |
|
|
Here are some examples. If you want to run a NetBSD/pmax kernel on an |
166 |
|
|
emulated DECstation machine, you would use a command line such as this: |
167 |
|
|
<pre> |
168 |
dpavlin |
12 |
$ <b>gxemul -e 3max -d pmax_diskimage.fs netbsd-pmax-INSTALL</b> |
169 |
dpavlin |
2 |
</pre> |
170 |
dpavlin |
20 |
|
171 |
|
|
<p>NOTE: For some emulation modes, such as the DECstation mode, you do |
172 |
|
|
<i>not</i> actually have to specify the name of the kernel, if the disk |
173 |
|
|
image is bootable! |
174 |
|
|
|
175 |
|
|
<p>It is possible to have more than one disk. For each -d argument, a disk |
176 |
dpavlin |
2 |
image is added; the first will be SCSI target 0, the second will be target 1, and so on, |
177 |
|
|
unless you specify explicitly which ID number the devices should have. |
178 |
|
|
<pre> |
179 |
dpavlin |
12 |
$ <b>gxemul -e 3max -d disk0.raw -d disk1.raw -d 5:disk2.raw netbsd-pmax-INSTALL</b> |
180 |
dpavlin |
2 |
</pre> |
181 |
|
|
Note: In the example above, disk2.raw will get scsi id 5. |
182 |
dpavlin |
20 |
|
183 |
|
|
<p>If a filename has a 'c' prefix, or ends with ".iso", then it is assumed to be |
184 |
dpavlin |
2 |
a CDROM device (this can be overridden with a 'd' prefix, to force a read/write disk). |
185 |
|
|
For example, the following command would start the emulator with two |
186 |
|
|
CDROM images, and one harddisk image: |
187 |
|
|
<pre> |
188 |
dpavlin |
12 |
$ <b>gxemul -e 3max -d image.iso -d disk0.img -d c:second_cdrom.img netbsd-pmax-INSTALL</b> |
189 |
dpavlin |
2 |
</pre> |
190 |
|
|
Usually, the device with the lowest id becomes the boot device. To override |
191 |
|
|
this, add a 'b' prefix to one of the devices: |
192 |
|
|
<pre> |
193 |
dpavlin |
12 |
$ <b>gxemul -e 3max -d rootdisk.img -d bc:install-cd.iso name_of_kernel</b> |
194 |
dpavlin |
2 |
</pre> |
195 |
|
|
If you have a physical CD-ROM drive on the host machine, say /dev/cd0c, you can |
196 |
|
|
use it as a CD-ROM directly accessible from within the emulator: |
197 |
|
|
<pre> |
198 |
dpavlin |
12 |
$ <b>gxemul -e 3max -d rootdisk.img -d bc:/dev/cd0c name_of_kernel</b> |
199 |
dpavlin |
2 |
</pre> |
200 |
|
|
It is probably possible to use harddisks as well this way, but I would not |
201 |
|
|
recommend it. |
202 |
dpavlin |
38 |
|
203 |
|
|
|
204 |
|
|
|
205 |
|
|
|
206 |
|
|
|
207 |
|
|
|
208 |
|
|
<p><br> |
209 |
|
|
<a name="tape_images"></a> |
210 |
|
|
<h3>How to start the emulator with tape images:</h3> |
211 |
|
|
|
212 |
dpavlin |
2 |
Using emulated tape drives is a bit more complicated than disks, because a |
213 |
|
|
tape can be made up of several "files" with space in between. The solution |
214 |
|
|
I have choosen is to have one file in the host's file system space for each |
215 |
|
|
tape file. The prefix for using tapes is 't', and the filename given is |
216 |
|
|
for the <i>first</i> file on that tape (number zero, implicitly). For |
217 |
|
|
files following file nr 0, a dot and the filenumber is appended to the |
218 |
|
|
filename. |
219 |
|
|
<p> |
220 |
|
|
As an example, starting the emulator with |
221 |
|
|
<pre> |
222 |
|
|
<b>-d t4:mytape.img</b> |
223 |
|
|
</pre> |
224 |
|
|
will cause SCSI id 4 to be a tape device, using the following file number |
225 |
|
|
to name translation scheme: |
226 |
|
|
<p> |
227 |
|
|
<center> |
228 |
|
|
<table border="0"> |
229 |
|
|
<tr> |
230 |
|
|
<td><b>File number:</b></td> |
231 |
|
|
<td><b>File name in the host's filesystem:</b></td> |
232 |
|
|
</tr> |
233 |
|
|
<tr> |
234 |
|
|
<td align="center">0</td> |
235 |
|
|
<td align="left">mytape.img</td> |
236 |
|
|
</tr> |
237 |
|
|
<tr> |
238 |
|
|
<td align="center">1</td> |
239 |
|
|
<td align="left">mytape.img.1</td> |
240 |
|
|
</tr> |
241 |
|
|
<tr> |
242 |
|
|
<td align="center">2</td> |
243 |
|
|
<td align="left">mytape.img.2</td> |
244 |
|
|
</tr> |
245 |
|
|
<tr> |
246 |
|
|
<td align="center">..</td> |
247 |
|
|
<td align="left">..</td> |
248 |
|
|
</tr> |
249 |
|
|
</table> |
250 |
|
|
</center> |
251 |
|
|
<p> |
252 |
|
|
If you already have a number of tape files, which should be placed on the |
253 |
|
|
same emulated tape, then you might not want to rename all those files. |
254 |
|
|
Use symbolic links instead (ln -s). |
255 |
|
|
<p> |
256 |
|
|
There is another advantage to using symbolic links for tape filenames: |
257 |
|
|
every time a tape is rewound, it is reopened using the filename given |
258 |
|
|
on the command line. By changing what the symbolic name points to, |
259 |
|
|
you can "switch tapes" without quiting and restarting the emulator. |
260 |
|
|
|
261 |
dpavlin |
38 |
<p> |
262 |
|
|
<font color="#ff0000">Note: Tape support is most likely very buggy, |
263 |
|
|
because it has not been tested much, and has probably also suffered |
264 |
|
|
from bit-rot by now.</font> |
265 |
dpavlin |
2 |
|
266 |
|
|
|
267 |
dpavlin |
20 |
|
268 |
|
|
|
269 |
|
|
|
270 |
dpavlin |
38 |
|
271 |
dpavlin |
2 |
<p><br> |
272 |
dpavlin |
38 |
<a name="disk_overlays"></a> |
273 |
|
|
<h3>How to use disk image overlays:</h3> |
274 |
|
|
|
275 |
|
|
This is most likely best understood by an example: |
276 |
|
|
|
277 |
|
|
<p><ul> |
278 |
|
|
<li>Install e.g. <a href="guestoses.html#netbsdcatsinstall">NetBSD/cats</a>. |
279 |
|
|
You will end up with a disk image called |
280 |
|
|
<tt>nbsd_cats.img</tt>. |
281 |
|
|
<p> |
282 |
|
|
<li>Running the following command will boot straight from the disk |
283 |
|
|
image, with no overlay images:<pre> |
284 |
|
|
<b>gxemul -XEcats -d nbsd_cats.img netbsd.aout-GENERIC.gz</b> |
285 |
|
|
|
286 |
|
|
</pre> |
287 |
|
|
<li>You may now create an overlay file, a corresponding map file, |
288 |
|
|
and start the emulator with the overlay image connected to |
289 |
|
|
the same (explicit) ID as the base disk image:<pre> |
290 |
|
|
<b>touch overlay.img overlay.img.map |
291 |
|
|
gxemul -XEcats -d 0:nbsd_cats.img -d V0:overlay.img netbsd.aout-GENERIC.gz</b> |
292 |
|
|
|
293 |
|
|
</pre> |
294 |
|
|
<li>Any changes to the filesystem you perform when using the overlay |
295 |
|
|
will only be written to that overlay. For example, to perform |
296 |
|
|
a "roll back", you can do the following:<pre> |
297 |
|
|
<b>rm -f overlay.img overlay.img.map |
298 |
|
|
touch overlay.img overlay.img.map</b> |
299 |
|
|
|
300 |
|
|
</pre> |
301 |
|
|
and then simply start the emulator again, with the newly created |
302 |
|
|
overlay image. |
303 |
|
|
</ul> |
304 |
|
|
|
305 |
|
|
<p>It is also possible to add multiple overlays. In that case, writes |
306 |
|
|
always go the the <i>last</i> added overlay. |
307 |
|
|
|
308 |
|
|
<p>GXemul uses Unix' way of supporting files with "holes", |
309 |
|
|
so even if <tt>ls -l overlay.img</tt> says that the overlay is several |
310 |
|
|
gigabytes large, <tt>du overlay.img</tt> should reveal that only the |
311 |
|
|
blocks that have actually been written to have been stored in the |
312 |
|
|
overlay, e.g.:<pre> |
313 |
|
|
<b>$ ls -l |
314 |
|
|
.. |
315 |
|
|
-rw-r--r-- 1 debug wheel 3072319488 Mar 24 11:59 nbsd_cats.img |
316 |
|
|
-rw-r--r-- 1 debug wheel 2465354 Mar 24 11:44 netbsd.aout-GENERIC.gz |
317 |
|
|
-rw-r--r-- 1 debug wheel 2930841600 Mar 24 14:02 overlay.img |
318 |
|
|
-rw-r--r-- 1 debug wheel 715538 Mar 24 14:02 overlay.img.map |
319 |
|
|
$ du overlay.img |
320 |
|
|
864 overlay.img |
321 |
|
|
</b> |
322 |
|
|
</pre> |
323 |
|
|
|
324 |
|
|
<p>The .map file is simply a raw bitmap telling which blocks of the |
325 |
|
|
overlay file that are in use. |
326 |
|
|
|
327 |
|
|
|
328 |
|
|
|
329 |
|
|
|
330 |
|
|
|
331 |
|
|
<p><br> |
332 |
dpavlin |
20 |
<a name="filexfer"></a> |
333 |
|
|
<h3>Transfering files to/from the guest OS:</h3> |
334 |
|
|
|
335 |
dpavlin |
32 |
If the emulated machine supports networking (see <a |
336 |
dpavlin |
40 |
href="networking.html#intro">this section</a> for more info), then the easiest |
337 |
|
|
way to transfer files is probably to use FTP or similar methods. |
338 |
dpavlin |
20 |
|
339 |
|
|
<p>There is another way of transfering files which works for any kind of |
340 |
|
|
emulated machine which supports disks (either SCSI or IDE). Any file can |
341 |
|
|
be supplied as a disk image. For example, consider the following:<pre> |
342 |
|
|
$ <b>gxemul -XEcats -d nbsd_cats.img -d archive.tar.gz netbsd-GENERIC</b> |
343 |
|
|
</pre> |
344 |
|
|
This will start NetBSD/cats with <tt>nbsd_cats.img</tt> as IDE master on |
345 |
|
|
controller 0 (wd0), and <tt>archive.tar.gz</tt> as IDE slave on controller |
346 |
|
|
0 (wd1). From inside NetBSD, it is now possible to extract the files using |
347 |
|
|
the following command:<pre> |
348 |
|
|
(inside emulated NetBSD/cats) |
349 |
|
|
# <b>tar zxvf /dev/wd1c</b> |
350 |
|
|
</pre> |
351 |
|
|
Don't worry if NetBSD complains about lack of disklabel; it doesn't |
352 |
|
|
matter. On some machines, NetBSD uses <tt>wd1d</tt> instead of |
353 |
|
|
<tt>wd1c</tt> for the entire disk. |
354 |
|
|
There is also a minor problem: reading the end of the disk image. If you |
355 |
|
|
experience problems untaring archives like this, then pad out the archive |
356 |
|
|
first with some zeroes. |
357 |
|
|
|
358 |
|
|
<p>Transfering files <i>out</i> from the emulated operating system to the |
359 |
|
|
host can be done the same way. First, prepare an empty archive file:<pre> |
360 |
|
|
$ <b>dd if=/dev/zero of=newarchive.tar bs=1024 count=1 seek=10000</b> |
361 |
|
|
</pre>This example created a 10 MB empty file. Then, start the emulator |
362 |
|
|
like this:<pre> |
363 |
|
|
$ <b>gxemul -XEcats -d nbsd_cats.img -d archive.tar netbsd-GENERIC</b> |
364 |
|
|
</pre> |
365 |
|
|
and transfer files by creating an archive directly onto the disk image:<pre> |
366 |
|
|
(inside emulated NetBSD/cats) |
367 |
|
|
# <b>tar cvf /dev/wd1c filenames</b> |
368 |
|
|
</pre> |
369 |
|
|
where filenames are the files or directories to transfer. |
370 |
|
|
|
371 |
|
|
|
372 |
|
|
|
373 |
|
|
|
374 |
|
|
|
375 |
|
|
<p><br> |
376 |
dpavlin |
2 |
<a name="largeimages"></a> |
377 |
|
|
<h3>How to extract large gzipped disk images:</h3> |
378 |
|
|
|
379 |
|
|
Unix filesystems usually support large files with "holes". Holes are |
380 |
|
|
zero-filled blocks that don't actually exist on disk. This is very |
381 |
|
|
practical for emulated disk images, as it is possible to create a very |
382 |
|
|
large disk image without using up much space at all. |
383 |
|
|
|
384 |
|
|
<p> |
385 |
|
|
Using gzip and gunzip on disk images can be <i>very</i> slow, as these |
386 |
|
|
files can be multiple gigabytes large, but this is usually necessary for |
387 |
|
|
transfering disk images over the internet. If you receive a gzipped disk |
388 |
|
|
image, say disk.img.gz, and run a naive |
389 |
|
|
<p> |
390 |
|
|
<pre> |
391 |
|
|
$ <b>gunzip disk.img.gz</b> |
392 |
|
|
</pre> |
393 |
|
|
<p> |
394 |
|
|
on it, you will not end up with an optimized file unless |
395 |
|
|
gunzip supports that. (In my experiments, it doesn't.) In plain English, |
396 |
|
|
if you type <b>ls -l</b> and the filesize is 9 GB, it will actually occupy |
397 |
|
|
9 GB of disk space! This is often unacceptable. |
398 |
|
|
<p> |
399 |
|
|
Using a simple tool which only writes blocks that are non-zero, a lot of |
400 |
|
|
space can be saved. Compile the program cp_removeblocks in the |
401 |
|
|
experiments/ directory, and type: |
402 |
|
|
<p> |
403 |
|
|
<pre> |
404 |
|
|
$ <b>gunzip -c disk.img.gz | cp_removeblocks /dev/stdin disk.img</b> |
405 |
|
|
</pre> |
406 |
|
|
|
407 |
|
|
<p> |
408 |
|
|
This will give you a disk.img which looks like it is 9 GB, and works like |
409 |
|
|
the real file, but the holes are not written out to the disk. (You can see |
410 |
|
|
this by running for example <b>du disk.img</b> to see the physical block |
411 |
|
|
count.) |
412 |
|
|
|
413 |
|
|
|
414 |
|
|
|
415 |
|
|
<p><br> |
416 |
|
|
<a name="userland"></a> |
417 |
|
|
<h3>Running userland binaries:</h3> |
418 |
|
|
|
419 |
dpavlin |
24 |
<font color="#ff0000">Note: This feature does not really work yet. |
420 |
|
|
It is currently disabled in stable release builds of the emulator.</font> |
421 |
dpavlin |
2 |
|
422 |
dpavlin |
12 |
<p>There is some skeleton code for running userland programs as well. This |
423 |
|
|
will not emulate any particular machine, but instead try to translate |
424 |
|
|
syscalls from e.g. NetBSD/pmax into the host's OS' syscalls. Right now, |
425 |
|
|
this is just a proof-of-concept, to show that it could work; there's lots |
426 |
|
|
of work left to do to make it actually run useful programs. |
427 |
|
|
|
428 |
dpavlin |
2 |
<p> |
429 |
|
|
|
430 |
|
|
<ul> |
431 |
|
|
<li><b>NetBSD/pmax:</b> |
432 |
|
|
<br> |
433 |
|
|
Running /bin/hostname or /bin/date and similarly trivial |
434 |
|
|
programs from the NetBSD/pmax distribution works:<pre> |
435 |
|
|
$ <b>gxemul -q -u netbsd/pmax pmax_bin_hostname</b> |
436 |
|
|
tab.csbnet.se |
437 |
|
|
$ <b>gxemul -q -u netbsd/pmax pmax_bin_date</b> |
438 |
|
|
Sun Jan 25 02:26:14 GMT 2004 |
439 |
|
|
$ <b>gxemul -q -u netbsd/pmax pmax_bin_sleep</b> |
440 |
|
|
usage: pmax_bin_sleep seconds |
441 |
|
|
$ <b>gxemul -q -u netbsd/pmax pmax_bin_sleep 5</b> |
442 |
|
|
$ <b>gxemul -q -u netbsd/pmax pmax_bin_sync</b> |
443 |
|
|
</pre> |
444 |
|
|
|
445 |
|
|
<p> |
446 |
|
|
<li><b>Ultrix:</b> |
447 |
|
|
<br> |
448 |
|
|
At least /bin/date and /bin/hostname work:<pre> |
449 |
|
|
$ <b>gxemul -q -u ultrix ultrix4_bin_date</b> |
450 |
|
|
UNIMPLEMENTED ultrix syscall 54 |
451 |
|
|
UNIMPLEMENTED ultrix syscall 62 |
452 |
|
|
Mon Feb 9 12:50:59 WET 2004 |
453 |
|
|
$ <b>gxemul -q -u ultrix ultrix4_bin_hostname</b> |
454 |
|
|
tab.csbnet.se |
455 |
|
|
</pre> |
456 |
|
|
|
457 |
dpavlin |
12 |
<!-- |
458 |
dpavlin |
2 |
<p> |
459 |
|
|
<li><b>NetBSD/powerpc:</b> |
460 |
|
|
<br> |
461 |
|
|
/bin/sync from NetBSD/macppc works, but probably not much else.<pre> |
462 |
|
|
$ <b>gxemul -v -u netbsd/powerpc netbsd-1.6.2-macppc-bin-sync</b> |
463 |
|
|
... |
464 |
|
|
[ sync() ] |
465 |
|
|
[ exit(0) ] |
466 |
|
|
cpu_run_deinit(): All CPUs halted. |
467 |
|
|
|
468 |
|
|
</pre> |
469 |
|
|
|
470 |
|
|
<p> |
471 |
|
|
<li><b>Linux/PPC64:</b> |
472 |
|
|
<br> |
473 |
|
|
The <a href="http://www-106.ibm.com/developerworks/library/l-ppc/#h13">64-bit Hello World assembly language example</a> |
474 |
|
|
on IBM's developerWorks pages runs:<pre> |
475 |
|
|
$ <b>ppc64-unknown-linux-as hello-ppc64.s -o hello-ppc64.o</b> |
476 |
|
|
$ <b>ppc64-unknown-linux-ld hello-ppc64.o -o hello-ppc64</b> |
477 |
|
|
$ <b>gxemul -q -u linux/ppc64 hello-ppc64</b> |
478 |
|
|
Hello, world! |
479 |
|
|
|
480 |
|
|
</pre> |
481 |
dpavlin |
12 |
--> |
482 |
dpavlin |
2 |
|
483 |
|
|
</ul> |
484 |
|
|
|
485 |
|
|
|
486 |
|
|
|
487 |
|
|
|
488 |
|
|
|
489 |
|
|
<p><br> |
490 |
|
|
<a name="promdump"></a> |
491 |
|
|
<h3>Using a PROM dump from a real machine:</h3> |
492 |
|
|
|
493 |
|
|
Raw PROM images from real machines can, in a few cases, be used in |
494 |
|
|
the emulator. ROM code is usually much more sensitive to correctness |
495 |
|
|
of the emulator than operating system kernels or userland programs |
496 |
|
|
are, so don't expect any PROM image to just magically work. |
497 |
|
|
|
498 |
|
|
|
499 |
|
|
<p> |
500 |
|
|
<h4>Dumping the PROM on a DECstation 5000/125:</h4> |
501 |
|
|
The image first needs to be extracted from the machine. There are |
502 |
|
|
several ways to do this. |
503 |
|
|
<ul> |
504 |
|
|
<li>Use hardware to read the PROM chip(s) directly. Not easy if you |
505 |
|
|
don't have such a hardware reader. |
506 |
|
|
<li>Copy the PROM memory range into a file, from a running |
507 |
|
|
operating system. You need a running OS, and it must |
508 |
|
|
have access to the PROM memory range. NetBSD, for example, |
509 |
|
|
doesn't allow that from userland. |
510 |
|
|
<li>Hook up a serial console and dump using the PROM's own dump |
511 |
|
|
command. |
512 |
|
|
</ul> |
513 |
|
|
<p> |
514 |
|
|
The easiest way is to hook up a serial console. The terminal must be |
515 |
|
|
able to capture output to a file. |
516 |
|
|
<p> |
517 |
|
|
These are approximately the commands that I used: |
518 |
|
|
<pre> |
519 |
|
|
>><b>cnfg</b> <i>Show machine configuration</i> |
520 |
|
|
|
521 |
|
|
>><b>printenv</b> <i>Show environment variables</i> |
522 |
|
|
|
523 |
|
|
>><b>setenv more 0</b> <i>This turns off the More messages</i> |
524 |
|
|
|
525 |
|
|
>><b>e -x 0xbfc00000:0xbfffffff</b> <i>Dump the PROM data</i> |
526 |
|
|
</pre> |
527 |
|
|
<p> |
528 |
|
|
Remember that DECstations are little endian, so if the dump data |
529 |
|
|
looks like this: |
530 |
|
|
<pre> |
531 |
|
|
bfc00000: 0x0bf0007e |
532 |
|
|
</pre> |
533 |
|
|
then the bytes in memory are actually 0x7e, 0x00, 0xf0, and 0x0b. |
534 |
|
|
<p> |
535 |
|
|
At 9600 bps, about 10KB can be dumped per minute, so it takes a while. |
536 |
|
|
Once enough of the PROM has been dumped, you can press CTRL-C to break out. |
537 |
|
|
Then, restore the more environment variable: |
538 |
|
|
<pre> |
539 |
|
|
>><b>setenv more 24</b> |
540 |
|
|
</pre> |
541 |
|
|
<p> |
542 |
|
|
Now, convert the data you just saved (little-endian words -> bytes), |
543 |
|
|
and store in a file. Let's call this file DECstation5000_125_promdump.bin. |
544 |
|
|
<pre> |
545 |
|
|
$ <b>decprom_dump_txt_to_bin DECstation5000_125_promdump.txt DECstation5000_125_promdump.bin</b> |
546 |
|
|
</pre> |
547 |
|
|
This binary image can now be used in the emulator: |
548 |
|
|
<pre> |
549 |
dpavlin |
12 |
$ <b>gxemul -e 3min -Q -M128 -q 0xbfc00000:DECstation5000_125_promdump.bin</b> |
550 |
dpavlin |
2 |
|
551 |
|
|
KN02-BA V5.7e |
552 |
|
|
?TFL: 3/scc/access (1:Ln1 reg-12: actual=0x00 xpctd=0x01) [KN02-BA] |
553 |
|
|
?TFL: 3/scc/io (1:Ln0 tx bfr not empty. status=0X 0) [KN02-BA] |
554 |
|
|
... |
555 |
|
|
--More--?TFL: 3/scsi/cntl (CUX, cause= 1000002C) |
556 |
|
|
>><b>?</b> |
557 |
|
|
? [cmd] |
558 |
|
|
boot [[-z #] [-n] #/path [ARG...]] |
559 |
|
|
cat SCRPT |
560 |
|
|
cnfg [#] |
561 |
|
|
d [-bhw] [-S #] RNG VAL |
562 |
|
|
e [-bhwcdoux] [-S #] RNG |
563 |
|
|
erl [-c] |
564 |
|
|
go [ADR] |
565 |
|
|
init [#] [-m] [ARG...] |
566 |
|
|
ls [#] |
567 |
|
|
passwd [-c] [-s] |
568 |
|
|
printenv [EVN] |
569 |
|
|
restart |
570 |
|
|
script SCRPT |
571 |
|
|
setenv EVN STR |
572 |
|
|
sh [-belvS] [SCRPT] [ARG..] |
573 |
|
|
t [-l] #/STR [ARG..] |
574 |
|
|
unsetenv EVN |
575 |
|
|
>><b>cnfg</b> |
576 |
|
|
3: KN02-BA DEC V5.7e TCF0 (128 MB) |
577 |
|
|
(enet: 00-00-00-00-00-00) |
578 |
|
|
(SCSI = 7) |
579 |
|
|
0: PMAG-BA DEC V5.3a TCF0 |
580 |
|
|
>><b>printenv</b> |
581 |
|
|
boot= |
582 |
|
|
testaction=q |
583 |
|
|
haltaction=h |
584 |
|
|
more=24 |
585 |
|
|
#=3 |
586 |
|
|
console=* |
587 |
|
|
osconsole=3 |
588 |
|
|
>> |
589 |
|
|
</pre> |
590 |
dpavlin |
14 |
|
591 |
|
|
<p><font color="#ff0000">(Note: at the moment, this doesn't work. |
592 |
|
|
I must have broken something when fixing something else, but this |
593 |
|
|
is what it looked like at the time.)</font> |
594 |
|
|
|
595 |
|
|
<p>During bootup, the PROM complains <i>a lot</i> about hardware failures. |
596 |
dpavlin |
2 |
That's because the emulator doesn't emulate the hardware well enough yet. |
597 |
dpavlin |
14 |
|
598 |
|
|
<p>The command line options used are: <tt>-e 3min</tt> for |
599 |
|
|
"DECstation 3min" (5000/1xx), <tt>-Q</tt> to supress the emulator's own PROM |
600 |
|
|
call emulation, <tt>-M128</tt> for 128MB RAM (because GXemul doesn't correctly |
601 |
dpavlin |
2 |
emulate memory detection well enough for the PROM to accept, so it will |
602 |
dpavlin |
14 |
always believe there is 128MB ram anyway), and <tt>-q</tt> to supress debug messages. |
603 |
|
|
The <tt>0xbfc00000</tt> in front of the filename tells GXemul that it is a raw |
604 |
dpavlin |
2 |
binary file which should be loaded at a specific virtual address. |
605 |
|
|
|
606 |
|
|
|
607 |
|
|
<p><br> |
608 |
|
|
<h4>Dumping the PROM on a SGI O2:</h4> |
609 |
|
|
|
610 |
|
|
The general ideas in this section applies to using ROM images from other |
611 |
dpavlin |
14 |
machines as well. I have also tried this on an SGI IP32 ("O2"), in addition |
612 |
|
|
to the DECstation. |
613 |
|
|
|
614 |
|
|
<p>For the O2, a suitable command to dump the prom memory range is |
615 |
dpavlin |
2 |
<pre> |
616 |
|
|
>> <b>dump -b 0xBFC00000:0xBFC80000</b> |
617 |
|
|
</pre> |
618 |
|
|
Make sure you capture all the output (via serial console) into a file, |
619 |
dpavlin |
14 |
and then run <tt>experiments/sgiprom_to_bin</tt> on the captured file. |
620 |
|
|
|
621 |
dpavlin |
2 |
<p> |
622 |
dpavlin |
14 |
|
623 |
|
|
<a href="sgi-o2-real.jpg"><img src="sgi-o2-real_small.jpg"></a> |
624 |
|
|
|
625 |
|
|
<a href="20050817-sgi-o2-success-7.png"><img src="20050817-sgi-o2-success-7_small.png"></a> |
626 |
|
|
|
627 |
|
|
<a href="20050817-sgi-o2-success-8.png"><img src="20050817-sgi-o2-success-8_small.png"></a> |
628 |
dpavlin |
2 |
|
629 |
dpavlin |
14 |
<p>The photo on the left is from the real machine. The other two are |
630 |
|
|
screenshots of the PROM running experimentally in GXemul, using <tt>-Y2</tt> |
631 |
|
|
framebuffer scaledown. |
632 |
dpavlin |
2 |
|
633 |
dpavlin |
14 |
<p>Normally during bootup, the IP32 PROM does a Power-On test which makes |
634 |
|
|
sure that the caches and other things are working properly. GXemul doesn't |
635 |
|
|
emulate all those things well enough for the tests to pass. The |
636 |
|
|
experimental screenshots above were taken with cache detection skipped |
637 |
|
|
manually. |
638 |
dpavlin |
2 |
|
639 |
dpavlin |
14 |
<p><font color="#ff0000"> |
640 |
|
|
In other words: don't expect this to work out-of-the-box with GXemul right |
641 |
|
|
now. It might work once I've added correct cache emulation.</font> |
642 |
dpavlin |
2 |
|
643 |
dpavlin |
14 |
<p>The command line used to start the emulator, once correct cache |
644 |
|
|
emulation has been implemented, would be something like <tt>gxemul -XQeo2 |
645 |
|
|
0xbfc00000:prom.bin</tt>. |
646 |
dpavlin |
2 |
|
647 |
dpavlin |
14 |
<p>The same caution applies when dealing with SGI PROMs as with |
648 |
|
|
DECstation PROMs: GXemul doesn't really emulate the hardware, it only |
649 |
|
|
"fakes" devices well enough to fool some things, primarily NetBSD, that |
650 |
|
|
it is emulating a real machine. ROM code is usually a <i>lot</i> more |
651 |
|
|
picky about the details. |
652 |
|
|
|
653 |
|
|
<p>The graphics used in the O2 is (as far as I know) undocumented. Combining |
654 |
|
|
some traces of info from how Linux/O2 draws to the screen with some |
655 |
dpavlin |
18 |
reverse-engineering of my own, I've implemented enough of the controller to |
656 |
dpavlin |
14 |
let the PROM draw rectangles and bitmaps, but not much more. The SCSI |
657 |
|
|
controller is not implemented yet either. |
658 |
|
|
|
659 |
|
|
|
660 |
|
|
|
661 |
|
|
|
662 |
dpavlin |
2 |
</p> |
663 |
|
|
|
664 |
|
|
</body> |
665 |
|
|
</html> |