/[gxemul]/trunk/doc/misc.html
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Diff of /trunk/doc/misc.html

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 12 by dpavlin, Mon Oct 8 16:18:38 2007 UTC revision 24 by dpavlin, Mon Oct 8 16:19:56 2007 UTC
# Line 4  Line 4 
4  <table border=0 width=100% bgcolor="#d0d0d0"><tr>  <table border=0 width=100% bgcolor="#d0d0d0"><tr>
5  <td width=100% align=center valign=center><table border=0 width=100%><tr>  <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">  <td align="left" valign=center bgcolor="#d0efff"><font color="#6060e0" size="6">
7  <b>Gavare's eXperimental Emulator:&nbsp;&nbsp;&nbsp;</b></font>  <b>Gavare's eXperimental Emulator:</b></font><br>
8  <font color="#000000" size="6"><b>Miscellaneous</b>  <font color="#000000" size="6"><b>Miscellaneous</b>
9  </font></td></tr></table></td></tr></table><p>  </font></td></tr></table></td></tr></table><p>
10    
11  <!--  <!--
12    
13  $Id: misc.html,v 1.50 2005/08/16 05:15:24 debug Exp $  $Id: misc.html,v 1.63 2006/05/22 04:53:52 debug Exp $
14    
15  Copyright (C) 2003-2005  Anders Gavare.  All rights reserved.  Copyright (C) 2003-2006  Anders Gavare.  All rights reserved.
16    
17  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
18  modification, are permitted provided that the following conditions are met:  modification, are permitted provided that the following conditions are met:
# Line 52  SUCH DAMAGE. Line 52  SUCH DAMAGE.
52          developing firmware, using GXemul</a>          developing firmware, using GXemul</a>
53    <li><a href="#compilercontruct">Using GXemul in compiler contruction courses</a>    <li><a href="#compilercontruct">Using GXemul in compiler contruction courses</a>
54    <li><a href="#disk">How to start the emulator with a disk image</a>    <li><a href="#disk">How to start the emulator with a disk image</a>
55      <li><a href="#filexfer">Transfering files to/from the guest OS</a>
56    <li><a href="#largeimages">How to extract large gzipped disk images</a>    <li><a href="#largeimages">How to extract large gzipped disk images</a>
57    <li><a href="#userland">Running userland binaries</a>    <li><a href="#userland">Running userland binaries</a>
58    <li><a href="#promdump">Using a PROM dump from a real machine</a>    <li><a href="#promdump">Using a PROM dump from a real machine</a>
# Line 72  the Internet. If you are interested in t Line 73  the Internet. If you are interested in t
73  reasons why networking is implemented in the emulator the way it currently  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">  is implemented, you might want to read the <a href="technical.html#net">
75  networking section in the technical documentation</a>.  networking section in the technical documentation</a>.
76  <p>  
77  The guest OS running inside the emulator uses a private IPv4 address, such  <p><font color="#ff0000">This is still experimental, hackish, and
78  as 10.0.0.1, and the emulator acts as a NAT-like gateway/firewall at IPv4  rather buggy. With NetBSD running as guest operating system, it mostly
79  address 10.0.0.254. To the outside world it will seem like it is the host's  works.</font>
80  OS that connects to other machines on the internet, not the guest OS.  
81  <p>  <p>When only one machine is being emulated, the following default values
82  <font color="#ff0000">NOTE: This is still experimental!  apply:<pre>
83  As of 2004-07-21, ARP + ICMP + UDP + TCP are emulated well enough to let          IPv4 address:                   10.0.0.1
84  NetBSD and OpenBSD install via ftp, and use the network for many normal          Netmask:                        255.0.0.0
85  activities, but not everything works yet.</font>          Gateway / default route:        10.0.0.254
86            Nameserver:                     10.0.0.254
87    </pre>
88    
89    <p>The emulated machine must of course have a NIC which is emulated
90    correctly. At the moment, the following NICs should work:
91    <ul>
92      <li><tt><b>ether</b></tt>, the "fake" experimental ethernet device
93            (documented <a href="experiments.html#expdevices_ether">here</a>)
94      <li><tt><b>le</b></tt>, Turbochannel Lance Ethernet, as used in
95            DECstation 5000/200 ("3max")
96      <li><tt><b>mec</b></tt>, the SGI O2's ethernet controller
97      <li><tt><b>dec21143</b></tt>, Digital's 21143 NIC (known as <tt>dc</tt>
98            in OpenBSD, or <tt>tlp</tt> in NetBSD)
99    </ul>
100    
101    <p>The emulator acts as a NAT-like gateway/firewall; to the outside world
102    it will seem like it is the host's OS that connects to other machines on
103    the internet, not the guest OS.
104    
105    
106    
# Line 94  activities, but not everything works yet Line 113  activities, but not everything works yet
113    
114  Is this a good idea?  The answer is yes and no, depending on the level of  Is this a good idea?  The answer is yes and no, depending on the level of
115  detail you need in your simulations. If you are developing an operating  detail you need in your simulations. If you are developing an operating
116  system or operating system kernel of your own, and wish to target  system or operating system kernel of your own, then the emulator can be a
117  MIPS-like systems in general, then the answer might be yes, for  complement to testing on real hardware.
 experimental purposes.  
118    
119  <p>Some examples of things that <i>don't</i> work, that you should keep in  <p>Important things to keep in mind:
 mind:  
120    
121  <ul>  <ul>
122          <li>Porting code to a specific machine mode, e.g. a Silicon Graphics          <li>Porting code to a specific machine mode, e.g. a Silicon Graphics
123          machine, using GXemul, will not necessarily cause the code to          machine, using GXemul, will not "magically" cause the code to
124          work on a real machine. Sometimes code works in GXemul which doesn't          work on a real machine. Sometimes code works in GXemul which doesn't
125          work on real hardware, sometimes it's the other way around.          work on real hardware, sometimes it's the other way around.
126    
# Line 111  mind: Line 128  mind:
128          <li>GXemul contains bugs, and many things are not yet implemented.          <li>GXemul contains bugs, and many things are not yet implemented.
129    
130          <p>          <p>
131          <li>I have only implemented devices in GXemul to the degree that          <li><b>Very important!</b> I have only implemented devices in GXemul
132          NetBSD, OpenBSD, etc, don't complain too much. One way to say it          to the degree that NetBSD, OpenBSD, Linux, etc don't complain too much.
133          is that the device implementations are "lazy hacks", based on the          <p>
134          assumption that the emulated OS is already developed and bug-free.          If you are developing a driver for a device which is emulated by
135          They are not intended to be used for development of new OS code,          GXemul, and your driver does not seem to be working, then the
136          so if you do that, then be prepared for bugs and inconsitencies.          probability of a bug in GXemul's implementation of the device is
137            very much higher than that of a bug in your driver.
138            <p>
139            The device implementations in GXemul are based on the assumption
140            that the emulated OS is already developed and bug-free. They are
141            not primarily intended to be used for development of new device
142            driver code in operating systems, so if you do that, then be
143            prepared for bugs and inconsitencies.
144          <p>          <p>
145          <li>CPU details in GXemul are usually wrong. If your code depends          <li>CPU details in GXemul are usually wrong. If your code depends
146          on, say, R10000 or MIPS64 specifics, chances are that GXemul will          on, say, R10000 or MIPS64 specifics, chances are that GXemul will
147          not be sufficient. Again, this was done as "lazy hacks", and pretty          not be sufficient. One example is different revisions of ISAs;
148          much assumes that the OS being emulated is already developed          64-bit MIPS instructions which should trigger an exception on a
149          and bug-free.          real 32-bit MIPS processor usually execute anyway in GXemul. Another
150            example is if userland code tries to access kernel memory; in some
151            cases there is protection against this, but not in all cases (to get
152            higher performance).
153          <p>          <p>
154          <li>Caches. There is no cache emulation in GXemul right now. Caches          <li>Caches. There is no cache emulation in GXemul right now. Caches
155          for R2000/R3000 are faked well enough to run NetBSD, Ultrix, etc          for R2000/R3000 are faked well enough to run NetBSD, Ultrix, etc
# Line 145  your code during development, but it sho Line 170  your code during development, but it sho
170  <h3>Using GXemul in compiler contruction courses:</h3>  <h3>Using GXemul in compiler contruction courses:</h3>
171    
172  If you are learning how to write a compiler, and wish to target a  If you are learning how to write a compiler, and wish to target a
173  realistic target platform, then MIPS (as emulated by GXemul)  realistic target platform, then MIPS or ARM (as emulated by GXemul)
174  might be a suitable choice.  might be suitable choices.
175    
176  <ul>  <ul>
177    <li><b>(+)</b>&nbsp;&nbsp;Your compiler needs to output real assembly    <li><b>(+)</b>&nbsp;&nbsp;Your compiler needs to output real assembly
178          language code, which the assembler (eg gas, the GNU assembler) can          language code, which the assembler (e.g. gas, the GNU assembler) can
179          then compile into object format, and then you need to link this          then compile into object format, and then you need to link this
180          into an executable image. This is much closer to how things work          into an executable image. This is much closer to how things work
181          in real life than running assembly language listings in a simulator          in real life than running assembly language listings in a simulator
182          (eg SPIM).          (e.g. SPIM).
183    <p>    <p>
184    <li><b>(-)</b>&nbsp;&nbsp;GXemul does not simulate out-of-order    <li><b>(-)</b>&nbsp;&nbsp;GXemul does not simulate out-of-order
185          execution, penalties related to instruction scheduling, or          execution, penalties related to instruction scheduling, or
# Line 182  emulated DECstation machine, you would u Line 207  emulated DECstation machine, you would u
207  <pre>  <pre>
208          $ <b>gxemul -e 3max -d pmax_diskimage.fs netbsd-pmax-INSTALL</b>          $ <b>gxemul -e 3max -d pmax_diskimage.fs netbsd-pmax-INSTALL</b>
209  </pre>  </pre>
210  <p>  
211  NOTE: For some emulation modes, such as the DECstation mode, you do  <p>NOTE: For some emulation modes, such as the DECstation mode, you do
212  <i>not</i> have to specify the name of the kernel, if the disk image is  <i>not</i> actually have to specify the name of the kernel, if the disk
213  bootable!  image is bootable!
214  <p>  
215  It is possible to have more than one disk. For each -d argument, a disk  <p>It is possible to have more than one disk. For each -d argument, a disk
216  image is added; the first will be SCSI target 0, the second will be target 1, and so on,  image is added; the first will be SCSI target 0, the second will be target 1, and so on,
217  unless you specify explicitly which ID number the devices should have.  unless you specify explicitly which ID number the devices should have.
218  <pre>  <pre>
219          $ <b>gxemul -e 3max -d disk0.raw -d disk1.raw -d 5:disk2.raw netbsd-pmax-INSTALL</b>          $ <b>gxemul -e 3max -d disk0.raw -d disk1.raw -d 5:disk2.raw netbsd-pmax-INSTALL</b>
220  </pre>  </pre>
221  Note: In the example above, disk2.raw will get scsi id 5.  Note: In the example above, disk2.raw will get scsi id 5.
222  <p>  
223  If a filename has a 'c' prefix, or ends with ".iso", then it is assumed to be  <p>If a filename has a 'c' prefix, or ends with ".iso", then it is assumed to be
224  a CDROM device (this can be overridden with a 'd' prefix, to force a read/write disk).  a CDROM device (this can be overridden with a 'd' prefix, to force a read/write disk).
225  For example, the following command would start the emulator with two  For example, the following command would start the emulator with two
226  CDROM images, and one harddisk image:  CDROM images, and one harddisk image:
# Line 266  you can "switch tapes" without quiting a Line 291  you can "switch tapes" without quiting a
291    
292    
293    
294    
295    
296    
297    <p><br>
298    <a name="filexfer"></a>
299    <h3>Transfering files to/from the guest OS:</h3>
300    
301    If the emulated machine supports networking (see
302    <a href="#networking">above</a>), then transfering files via FTP is
303    probably easiest.
304    
305    <p>There is another way of transfering files which works for any kind of
306    emulated machine which supports disks (either SCSI or IDE). Any file can
307    be supplied as a disk image. For example, consider the following:<pre>
308            $ <b>gxemul -XEcats -d nbsd_cats.img -d archive.tar.gz netbsd-GENERIC</b>
309    </pre>
310    This will start NetBSD/cats with <tt>nbsd_cats.img</tt> as IDE master on
311    controller 0 (wd0), and <tt>archive.tar.gz</tt> as IDE slave on controller
312    0 (wd1). From inside NetBSD, it is now possible to extract the files using
313    the following command:<pre>
314            (inside emulated NetBSD/cats)
315            # <b>tar zxvf /dev/wd1c</b>
316    </pre>
317    Don't worry if NetBSD complains about lack of disklabel; it doesn't
318    matter. On some machines, NetBSD uses <tt>wd1d</tt> instead of
319    <tt>wd1c</tt> for the entire disk.
320    There is also a minor problem: reading the end of the disk image. If you
321    experience problems untaring archives like this, then pad out the archive
322    first with some zeroes.
323    
324    <p>Transfering files <i>out</i> from the emulated operating system to the
325    host can be done the same way. First, prepare an empty archive file:<pre>
326            $ <b>dd if=/dev/zero of=newarchive.tar bs=1024 count=1 seek=10000</b>
327    </pre>This example created a 10 MB empty file. Then, start the emulator
328    like this:<pre>
329            $ <b>gxemul -XEcats -d nbsd_cats.img -d archive.tar netbsd-GENERIC</b>
330    </pre>
331    and transfer files by creating an archive directly onto the disk image:<pre>
332            (inside emulated NetBSD/cats)
333            # <b>tar cvf /dev/wd1c filenames</b>
334    </pre>
335    where filenames are the files or directories to transfer.
336    
337    
338    
339    
340    
341  <p><br>  <p><br>
342  <a name="largeimages"></a>  <a name="largeimages"></a>
343  <h3>How to extract large gzipped disk images:</h3>  <h3>How to extract large gzipped disk images:</h3>
# Line 310  count.) Line 382  count.)
382  <a name="userland"></a>  <a name="userland"></a>
383  <h3>Running userland binaries:</h3>  <h3>Running userland binaries:</h3>
384    
385  <font color="#ff0000">Note: This does not really work yet.</font>  <font color="#ff0000">Note: This feature does not really work yet.
386    It is currently disabled in stable release builds of the emulator.</font>
387    
388  <p>There is some skeleton code for running userland programs as well. This  <p>There is some skeleton code for running userland programs as well. This
389  will not emulate any particular machine, but instead try to translate  will not emulate any particular machine, but instead try to translate
# Line 480  This binary image can now be used in the Line 553  This binary image can now be used in the
553           osconsole=3           osconsole=3
554          >>          >>
555  </pre>  </pre>
556  <i>(Note: at the moment, this doesn't work. I must have broken something when  
557  fixing something else, but this is what it looked like at the time.)</i>  <p><font color="#ff0000">(Note: at the moment, this doesn't work.
558  <p>  I must have broken something when fixing something else, but this
559  During bootup, the PROM complains <i>a lot</i> about hardware failures.  is what it looked like at the time.)</font>
560    
561    <p>During bootup, the PROM complains <i>a lot</i> about hardware failures.
562  That's because the emulator doesn't emulate the hardware well enough yet.  That's because the emulator doesn't emulate the hardware well enough yet.
563  <p>  
564  The command line options used are: -e 3min for  <p>The command line options used are: <tt>-e 3min</tt> for
565  "DECstation 3min" (5000/1xx), -Q to supress the emulator's own PROM  "DECstation 3min" (5000/1xx), <tt>-Q</tt> to supress the emulator's own PROM
566  call emulation, -M128 for 128MB RAM (because GXemul doesn't correctly  call emulation, <tt>-M128</tt> for 128MB RAM (because GXemul doesn't correctly
567  emulate memory detection well enough for the PROM to accept, so it will  emulate memory detection well enough for the PROM to accept, so it will
568  always believe there is 128MB ram anyway), and -q to supress debug messages.  always believe there is 128MB ram anyway), and <tt>-q</tt> to supress debug messages.
569  The 0xbfc00000 in front of the filename tells GXemul that it is a raw  The <tt>0xbfc00000</tt> in front of the filename tells GXemul that it is a raw
570  binary file which should be loaded at a specific virtual address.  binary file which should be loaded at a specific virtual address.
571    
572    
# Line 499  binary file which should be loaded at a Line 574  binary file which should be loaded at a
574  <h4>Dumping the PROM on a SGI O2:</h4>  <h4>Dumping the PROM on a SGI O2:</h4>
575    
576  The general ideas in this section applies to using ROM images from other  The general ideas in this section applies to using ROM images from other
577  machines as well. Besides DECstation, I've also tried this on an SGI IP32  machines as well. I have also tried this on an SGI IP32 ("O2"), in addition
578  ("O2").  to the DECstation.
579  <p>  
580  For the O2, a suitable command to dump the prom memory range is  <p>For the O2, a suitable command to dump the prom memory range is
581  <pre>  <pre>
582          &gt;&gt; <b>dump -b 0xBFC00000:0xBFC80000</b>          &gt;&gt; <b>dump -b 0xBFC00000:0xBFC80000</b>
583  </pre>  </pre>
584  Make sure you capture all the output (via serial console) into a file,  Make sure you capture all the output (via serial console) into a file,
585  and then run experiments/sgiprom_to_bin on the captured file.  and then run <tt>experiments/sgiprom_to_bin</tt> on the captured file.
 <p>  
 (2005-01-16: The emulator doesn't really emulate the IP32 well enough to  
 actually run the PROM image without using special hacks, but it might do  
 so some time in the future.)  
586    
587    <p>
588    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
589    <a href="sgi-o2-real.jpg"><img src="sgi-o2-real_small.jpg"></a>
590    &nbsp;&nbsp;&nbsp;
591    <a href="20050817-sgi-o2-success-7.png"><img src="20050817-sgi-o2-success-7_small.png"></a>
592    &nbsp;&nbsp;&nbsp;
593    <a href="20050817-sgi-o2-success-8.png"><img src="20050817-sgi-o2-success-8_small.png"></a>
594    
595    <p>The photo on the left is from the real machine. The other two are
596    screenshots of the PROM running experimentally in GXemul, using <tt>-Y2</tt>
597    framebuffer scaledown.
598    
599    <p>Normally during bootup, the IP32 PROM does a Power-On test which makes
600    sure that the caches and other things are working properly. GXemul doesn't
601    emulate all those things well enough for the tests to pass. The
602    experimental screenshots above were taken with cache detection skipped
603    manually.
604    
605    <p><font color="#ff0000">
606    In other words: don't expect this to work out-of-the-box with GXemul right
607    now. It might work once I've added correct cache emulation.</font>
608    
609    <p>The command line used to start the emulator, once correct cache
610    emulation has been implemented, would be something like <tt>gxemul -XQeo2
611    0xbfc00000:prom.bin</tt>.
612    
613    <p>The same caution applies when dealing with SGI PROMs as with
614    DECstation PROMs: GXemul doesn't really emulate the hardware, it only
615    "fakes" devices well enough to fool some things, primarily NetBSD, that
616    it is emulating a real machine. ROM code is usually a <i>lot</i> more
617    picky about the details.
618    
619    <p>The graphics used in the O2 is (as far as I know) undocumented. Combining
620    some traces of info from how Linux/O2 draws to the screen with some
621    reverse-engineering of my own, I've implemented enough of the controller to
622    let the PROM draw rectangles and bitmaps, but not much more. The SCSI
623    controller is not implemented yet either.
624    
625    
626    

Legend:
Removed from v.12  
changed lines
  Added in v.24

  ViewVC Help
Powered by ViewVC 1.1.26