The collected works of dexter1 - Page 1

Just my 2 cents for this matter:

gcc can be troublesome at times, so if you're fortunate enough to have the MIPSpro compilers, i usually try them before resorting to GNU. Typical make problems are incorrect environment flags like CC CFLAGS CXX CXXFLAGS, sometimes CPPFLAGS. Setting them to
CFLAGS="-I/usr/freeware/include -L/usr/freeware/lib"
Also some compiles break on national language support so a --disable-nls will fix that. Check also the LD_LIBRARY_PATH : mine is /usr/local/lib:/usr/freeware/lib:/usr/freeware/lib32

Beware however the tricky pointer arithmetic which is allowed in gcc, but MIPSpro barfs loudly with messages like:
cc-1133 cc: ERROR File = barf.c, Line = 751
Expression must be a modifiable lvalue.

(UINT32*)dst2 += 3;
so i changed it into dst2=(UINT32*)dst1 + 3;

That was my trick to compile Xmame :)

see ya
dexter1 wrote:
cc-1133 cc: ERROR File = barf.c, Line = 751
Expression must be a modifiable lvalue.

(UINT32*)dst2 += 3;
so i changed it into dst2=(UINT32*)dst1 + 3;

That should have been dst2=(UINT32*)dst2 + 3;

wuh, my first post and already a typo...
Well then,

My real name is Frank, and together with Fem i live in a small appartment in a suburb of The Hague, The Netherlands.
My job is being Yet Another SysAdm at Delft University of Technology. I was planning to go for a PhD in Groningen from 1994-1998, but switched to commercial ICT (Logica). That being too hectic for my way of life i decided to go back to University and do what i think i do best, helping people with Unix machines and scientific code.

I started Unix on an HP/Apollo, but in Groningen i quickly found a fun research group, and they had SGI machines, whoah!
Now i am sysadmin in a theoretical mechanics group, consisting of about 10 Indy's, which are now slowly being replaced by Linux machines, 3 O2's, a Indigo2 Impact, 3 beautiful Indigo1's :) , and 5 Origin 200 as our main servers and numbercrunchers. And yes a horde of Windoze machines, which i spend way to much time on...
Let me add my 2 cents with this thread.

From a professional point of view, i have to work with IRIX, because all my file and calculation servers run it. And let me say what a blessing that is. These machines run for a year straight without reboots. Stable OS, critical compilers (not gcc which swallows everything) and the only machines in the university who can pump full bandwith through fibercable or gigabit copper.
In general the SGI machines have excellent design and an OS which is custom made for these machines, thereby adding to the stablility.

Why you or anyone else should run SGI/IRIX? I don't know. Some people do it because they want to run Maya. Some people do it becuse they are fed up with redhat 8.x (i am fed up with that). Some people think it's cool to own and run such a machine (and cool it is ;) ). But lots of people run it because it is just plain simple fun:
Hey! You can point and click! What a cool gui! Wow OpenGL in hardware! What a neat GUI debugger (CVD)! How stable!

Do i run it at home? Not yet. I am happy with my 2.4 GHz win98/Linux (slackware) machine and play DiabloII and run GIMP. but an SGI machine as a fileserver? sure :)

Plus, and that is maybe my biggest reason i (still) use these machines, they are excellent in learning and understanding Unix, with good docs, good utils (xdiff should be on all unix machines)

But this is just my view... :)
Hello all,

For all of you Retrogamers, i finally have compiled Xmame 0.65.1 arcade machine emulator and Stella 1.3 Atari 2600 Emulator.

It appears that a bug crept in xmame 0.62.1 preventing it from compiling on SGI machines. That's why you see only xmame 0.61.1 on freeware...
I fixed that and Xmame is running happily. The fix is posted on the Xmame mailinglist, so future builds should be ok.

Stella is a different beast. It only emulates the good ol' Atari VCS 2600, but does so almost perfectly, even Pitfall II. Be sure to get the SDL library from freeware, otherwise sound is not available.

Check out the screenshot with Stella and Xmame in action:

I will give the tardist to Nekonoko in the near future, but i have to do lots of shit before i get around to it...
[Do you have any idea what type of system I would need to properly run the Neo Geo games? What about CPS2 games are they supported yet?

I tested XMame with some NeoGeo roms like PuzzleBobble2 and even with the fastest optimisation (i had to compile this on an Origin 200) i only got about 8-9 frames/s on my R5K180MHz O2. So i infer you should have at least an R10K and 192 MB for the NeoGeo stuff.

There may be a way to accelerate this BTW, i only compiled XMame with the MITSHM extensions, but looks like you can build it with OpenGL as well. I'll try that next week and i should have the results at the end of the week. Don't forget that mame is a very general Emulator. For selected consoles/systems, like NeoGeo, SNES, CPS2, there are other emulators, that are smaller, but faster. Look at for an overview.

For the CPS2 games, i found something on:
Looks like this is going to work. ROM's are probably unfindable :(

Are there any NES emulators for IRIX?

No compiled ones as yet, but it shouldn't be hard to port fwNES to IRIX. See also:
It seems that the NES is the console with the largest number of emulators...

What about controller support?

Hah! Well, get yourself a Fuel, that has USB on it and hook up your microsoft Sidewinder :twisted:
I don't know. Guess this is very system specific, but there was something at:
which looks like some abandoned page about controllers on PC, damn.

Hmmm, ok. My Girlie is not yet back from shopping yet, so lemme see what i can dig up using google:

SMS: Standard DB9 game controllers
MegaDrive: pinouts at: ... roller.asp
N64/PS2: Damn, usb :(

A nice hobby project for someone who has a year to spare, to hook all these beasts up to an O2/Octane/Indigo2...
Hello Neko,

Sorry for not replying earlier. My life at this moment is extremly hectic, with all the construction going on at my house, and work being an absolute bitch to me.

The latest version of XMame 0.66.2 compiles cleanly on SGI's with MIPSPro compilers, thanks to the project team honoring my patch requests. Only for the OpenGL port i had to add an extra GL header, but the GLport is severely unstable and cuts performance in half, compared to the X11 version.

By all means, try to compile it yourself, but beware: only X11 works, with MITSHM but without XV !
I will finalize an optimised build of XMame as soon as i get aroud to do it, wich will be in a couple of weeks. Oh and if you're at it , build GXMame as well. It's a great frontend to XMame.
Even though most of the ROM sites have been closed for free download, please do check out

Which gives info and links to some sites who offer a few ROMS for download.

Note however, that obtaining a full 0.66 ROM set that way is extremely difficult, because some roms are incredibly huge or hard to find or have been altered. :(

For people with older ROM sets check out this site:

It does not offer full ROM's for download, but only the changes from previous ROMsets to newer Mame releases.

Enjoy :)
Here at university, i can't do my own DNS, so they all have lame names as mech046, mech101 etc :(

In the early university days, workstations had names of small moons/planets, like Mimas, Pluto, Dione, Rhea and the big fileservers had names like Titan and Atlas. A fitting combination IMO.

SGI holland uses a rather convoluted naming scheme: first the decide if the machine is used for presales, support or demonstration, and then name the machine after a dutch city, which the first letter a P for Presales, an S for support etc. So they had in their demo room 8 O2's with names like Delft, DenHaag, Diemen etc...

In these times a serial connection to your Octane could he very helpful as well :)
Hello all,

I'll hereby give you my wacom file from /usr/lib/X11/input/config

With this config file i can remove my mouse from my desk and use the wacom tablet, an Intuos artpad II, as system mouse on my O2.

You need eoe.sw.optinput on the IRIX overlay's installed to use it and note that this doesn't enable your intuos 2 tablets, you'll need a separate driver for that...
Also the serial device manager needs to know you're using a wacom on a serial port.

Code: Select all

device_init {
left        "10160"
right       "10160"
top         "10160"
bottom      "10160"

x_init {
scale                  "fit"
scalewhich             "none"
scalex                 "1024/30480"
scaley                 "1024/30480"
scale0                 "1024/30480"
scale1                 "1024/30480"
pushpointer            "on"
autostart              "on"
relative               "on"
dblclickstylus         "1,25"
dblclickpuck           "1,25"
tipbutton              "3"
sidebutton             "1"
t+sbutton              "0"
ssbutton               "2"
eraserbutton           "2"
pressuremode           "on"
tippressure            "30"
eraserpressure         "30"
pressurecurve          "0,62,126,190,254"
puckbtn1               "2"
puckbtn2               "1"
puckbtn3               "4"
puckbtn4               "3"
wacpushx               "0"
wacpushy               "1"
wacpushp               "2"
wacpushxt              "5"
wacpushyt              "6"
wacpushz               "7"
invert                 "!xy0123456789"
Welcome Oddoh!

We have succesfully compiled xmame 0.65-1 and 0.66-1 (see the tardists Neko has made), and at this moment i'm trying to build xmess 0.66-1.
Considering the build problems with previous xmame/xmess releases have now been ironed out, this should not be hard. I'll let you know in two hours (!) if my Origin has spit out something useful...

Oh and xmame/xmess 0.68 is out and winmame 0.69 has just been released today...
Xmess compile fails with this statement:

Compiling src/cpu/m6805/m6805.c ...
cc-1143 cc: ERROR File = src/cpu/m6805/6805ops.c, Line = 75
Declaration is incompatible with "void bset(char *, bitnum_t)" (declared at
line 517 of "/usr/include/unistd.h").

INLINE void bset (UINT8 bit)

So i have to fix this... Hold on
Found it, it seems to be a header clash.

I have made a patch, recompiled the 0.67.2 binary and i have tested this with a vectrex image and seems to work fine.
I'll try to contact the xmame/xmess people about this so that we can implement this in xmame 0.69.1

If you have the MIPSPro compilers you can build it yourself with the following patch:

--- osdutils.h.old Sun Apr 6 23:41:28 2003
+++ osdutils.h Mon May 26 16:58:27 2003
@@ -1,7 +1,9 @@
#include <sys/stat.h>
#include <sys/types.h>
#include <fcntl.h>
+#ifndef __ARCH_irix
#include <unistd.h>

#define strcmpi strcasecmp
#define strncmpi strncasecmp

Patch it against xmame-0.67.2, enable xmess in the makefile, disable xv, and enable mit-shm

That should do the trick. Enjoy :)
Hi neko,

Thanks for the suggestion for #ifndef __ARCH_irix_al . I missed that one.

Nono the INLINE should be static :( Only then i can get the code to compile...

On my Origin i can compile xmess.x11 just fine, tho it needs 1.5 Gig to do so. I have just finished compiling xmess.x11 version 0.69.1 (it's out today).

Shall i ftp it to you, neko, so that you can make a tardist of it?

Oh and what should we make the binary? mips3 or mips4? Gerneral build or optimise for IP30?
You're right, they're broken....

I'll try again with both 0.67.2 and 0.69.1. If 0.69.1 fails, i have to backtrack the changes to the code :( ((

Neko, please remove the binaries from incoming and i'll make 0.67.2 builds of both, and i will upload them before noon (it's 0900 now)

Sorry for the mishap
i'll rebuild xmame also for 0.67.2 so we'll have that upgraded as well

Currently i'm investigating the bus error on 0.69.1. Seems someone has put broken code into the xmamerc/xmessrc handler.

Must be a Linux guy :twisted:
Currently i am testing a selfmade patch against issues in the VICE 1.11 code. Together with a CFLAGS setenv i believe i can compile VICE with MIPSPro (IRIX 6.5.19) after all. The code is very incoherent, sometimes ANSI, sometimes POSIX, sometimes BSD. I heard from lisp that some parts go waaaaay back....
Hi all,

I've just completed the patch and subsequent build tests of VICE 1.11 ( ). It seems to work fine, except for sound, which i don't know if it was broken in the first place.

Do the following if you want to compile it yourself with MIPSPro 7.3x
First, download and unpack the source at: ... .11.tar.gz

To apply the patch to the source listed below, save it to /tmp/vicesgi.patch and do:

Code: Select all

cd vice-1.11
patch -p0 < /tmp/vicesgi.patch

Then do:

Code: Select all

setenv CC cc
setenv CXX CC
setenv CFLAGS '-I/usr/freeware/include -L/usr/freeware/lib32 -mips4 -Ofast -IPA -ansi -pedantic -D_POSIX_SOURCE -D_BSD_SOURCE -D__EXTENSIONS__ -D_BSD_TYPES'

...and cross your fingers. After that a

Code: Select all

gmake install
as root should install everything.

People who are not blessed with MIPSPro 7.3x can use 7.2.1, provided they remove -Ofast -IPA from the CFLAGS and delete LDFLAGS alltogether

So here comes the patch. It is mothafoqqing uglee, probably breaks other (linux?) builds, makes your Indy explode and empties your fridge. In that order. And oh yeah, it needs more work...

Code: Select all

--- src/arch/unix/readline/   Sun Oct  7 23:05:39 2001
+++ src/arch/unix/readline/complete.c   Mon Jun  2 09:42:09 2003
@@ -78,7 +78,7 @@
*avp = av = new;

-        if ((av[ac] = strdup(p)) == NULL) {
+        if ((av[ac] = (char *)strdup(p)) == NULL) {
if (ac == 0)
@@ -107,18 +107,18 @@
char        *fpart;

if ((fpart = strrchr(path, '/')) == NULL) {
-        if ((dpart = strdup(DOT)) == NULL)
+        if ((dpart = (char *)strdup(DOT)) == NULL)
return -1;
-        if ((fpart = strdup(path)) == NULL) {
+        if ((fpart = (char *)strdup(path)) == NULL) {
return -1;
else {
-        if ((dpart = strdup(path)) == NULL)
+        if ((dpart = (char *)strdup(path)) == NULL)
return -1;
dpart[fpart - path] = '\0';
-        if ((fpart = strdup(++fpart)) == NULL) {
+        if ((fpart = (char *)strdup(++fpart)) == NULL) {
return -1;

--- src/plus4/   Thu Jan 16 20:23:37 2003
+++ src/plus4/ted.c   Tue Jun  3 13:46:21 2003
@@ -134,7 +134,7 @@
old_maincpu_clk += num;

-inline void ted_delay_clk(void)
+void ted_delay_clk(void)
CLOCK diff;

@@ -150,7 +150,7 @@

-inline void ted_handle_pending_alarms(int num_write_cycles)
+void ted_handle_pending_alarms(int num_write_cycles)

--- src/plus4/   Sat Oct 12 18:45:02 2002
+++ src/plus4/ted-fetch.c   Tue Jun  3 13:45:45 2003
@@ -39,7 +39,7 @@

/* Emulate a matrix line fetch, `num' bytes starting from `offs'.  This takes
care of the 10-bit counter wraparound.  */
-inline void ted_fetch_matrix(int offs, int num)
+void ted_fetch_matrix(int offs, int num)
BYTE *p;
int start_char;
@@ -60,7 +60,7 @@
memcpy(ted.cbuf, ted.cbuf_tmp, TED_SCREEN_TEXTCOLS);

-inline void ted_fetch_color(int offs, int num)
+void ted_fetch_color(int offs, int num)
int start_char;
int c;

--- src/drive/   Sat Jan 18 17:31:29 2003
+++ src/drive/drivecpu.c   Tue Jun  3 13:20:06 2003
@@ -322,7 +322,7 @@

-inline void drive_cpu_wake_up(drive_context_t *drv)
+void drive_cpu_wake_up(drive_context_t *drv)
/* FIXME: this value could break some programs, or be way too high for
others.  Maybe we should put it into a user-definable resource.  */
@@ -333,7 +333,7 @@

-inline void drive_cpu_sleep(drive_context_t *drv)
+void drive_cpu_sleep(drive_context_t *drv)
/* Currently does nothing.  But we might need this hook some day.  */

--- src/drive/   Mon Jan 20 19:44:22 2003
+++ src/drive/drive.c   Tue Jun  3 13:53:44 2003
@@ -527,7 +527,7 @@
The return value corresponds to bit#7 of VIA2 PRB. This means 0x0
is returned when sync is found and 0x80 is returned when no sync
is found.  */
-inline BYTE drive_sync_found(drive_t *dptr)
+BYTE drive_sync_found(drive_t *dptr)
BYTE val = dptr->GCR_track_start_ptr[dptr->GCR_head_offset];

@@ -685,7 +685,7 @@

/* Return the write protect sense status. */
-inline BYTE drive_write_protect_sense(drive_t *dptr)
+BYTE drive_write_protect_sense(drive_t *dptr)
/* Set the write protection bit for the time the disk is pulled out on
detach.  */

--- src/vicii/   Sat Jan 11 17:44:55 2003
+++ src/vicii/vicii-irq.c   Mon Jun  2 16:27:42 2003
@@ -35,7 +35,7 @@
#include "viciitypes.h"

-inline void vicii_irq_set_line(void)
+void vicii_irq_set_line(void)
if (vic_ii.irq_status & vic_ii.regs[0x1a]) {
vic_ii.irq_status |= 0x80;

--- src/vicii/   Thu Jan 16 19:54:47 2003
+++ src/vicii/vicii.c   Mon Jun  2 16:33:04 2003
@@ -172,7 +172,7 @@
old_maincpu_clk += num;

-inline void vicii_delay_clk(void)
+void vicii_delay_clk(void)
#if 0
CLOCK diff;
@@ -190,7 +190,7 @@

-inline void vicii_handle_pending_alarms(int num_write_cycles)
+void vicii_handle_pending_alarms(int num_write_cycles)
if (vic_ii.viciie != 0)

--- src/vicii/     Tue Jun  3 16:32:01 2003
+++ src/vicii/vicii-fetch.c       Mon Jun  2 16:33:48 2003
@@ -44,7 +44,7 @@

/* Emulate a matrix line fetch, `num' bytes starting from `offs'.  This takes
care of the 10-bit counter wraparound.  */
-inline void vicii_fetch_matrix(int offs, int num, int num_0xff)
+void vicii_fetch_matrix(int offs, int num, int num_0xff)
int start_char;
int c;

--- src/monitor/   Sat Dec 28 21:17:52 2002
+++ src/monitor/mon_parse.y   Mon Jun  2 09:33:59 2003
@@ -38,7 +38,7 @@
#else  /* Not HAVE_ALLOCA_H.  */
#if !defined(_AIX) && !defined(WINCE)
#ifndef _MSC_VER
-extern char *alloca();
+/* extern char *alloca(); */
#define alloca(n)   _alloca(n)
#endif  /* MSVC */
No, because the filesystem type needed to mount smbfs is not supported yet under IRIX. Smbmount doesn't seem to be supported under IRIX as well.

What you need is Sharity, an SMB client with GUI which can mount your SMB shares on IRIX machines. I use it on my O2 for two years now with great succes.

Freeware has a Sharity client and without the license you can use it, although support for shares will be limited to three levels deep.
Getting back to compiling xmess/xmame 0.69.1, it seems that 0.68.1 already is affected with the 'bus error'.

Wich leaves me less patchfiles to sift through, although it still is 3 MB bzip2 compressed...

I have CVD tho and have compiled xmess with -g so i can trace out what's going on.

To be continued.
Lisp suggests that a setenv CXXFLAGS '-mips4 -Ofast' should also be applied to speed up execution.

Oh and for all the folks who had a hard time extracting the patch from the forum message, i've put it here:
Thanks to guys at the xmame development list, the bus error on IRIX for xmame/xmess-0.69.1 has been found!

in src/mame.h they merged the wrong declaration for variable crc_only...

it should be: int crc_only;

instead of char.
Hm, i see what i can do this weekend. I'll have it definately by wednesday, so watch your FTP server over the weekend, neko.

After i finished construction in my house, i'll get on it :?
Hah! Too hot to do much construction anyway, so i cooked xmame and xmess and both are now on Neko's ftp server in incoming.

They are binaries from 0.69.1 and are baked -mips4 plus the -Ofast optimisation.

Neko, can you give them a spin, before making a tardist? I'm at home without a SGI on my desk yet...

Thanks for the Tardists Neko!

Sum][one, in order to play arcade machines with xmame you need a collection of roms each in a zipfile to play the arcade. For instance a rom file contains about ten pacman roms and to run it just type

xmame pacman

if all is well it will fire up pacman. press 5 for inserting coins and 1 or 2 for 1 or 2 player games

NES, SNES, 2600, Vectrex, C64, MSX and more are all handles by xmess. I myself am interested in vectrex, so my commandline looks like:

xmess vectrex -cart webwars.bin

To play the cartridge webwars on the vectrex system.
i can recommend GXMame, but i have only used that at home for some time now.

BTW xmame-0.70.1 will be out soon. Let's hope the builds are ok this time
I'll look into this tomorrow. I will attempt both xmame and xmess 0.70.1 builds and try the SDL as well, tho you're not gaining anything with respect to sound, because IRIX_AL is built in.
Ummm, i just built xmame 0.70.1 with CFLAGS = -n32 -mips4 -O2 -OPT:Olimit=0 and corresponding LD flags and it seems to run fine over here.
I impressed my colleagues with tempest and Strikers 1945 at 5 fps :)

I'll attempt a fully optimised build tomorrow, because this machine will be serviced tonight. It's not impossible the fully optimised build is broken in 0.70, tho i think we can say 0.70.1 irix build is safe...
I didn't mention that. I use MIPSPro which seems fine for building xmame.

Neko, since channelbundling has succeeded on "that" O200 (more or less, we're still ironing out the details) i can use the machine to build the xmame/xmess 0.70.1 binaries.

You'll see them on your ftp server soon...
I've googled a bit and have seen reports on IRIX 5.1 sep '93 ... du&rnum=54

This would seem as the earliest IRIX supporting the Indy
Hi all,

These are the release-notes for patch 5086, which i find quite interesting, to say the least. I found this while scourging for 6.5.21 info. Oh, and NO , i don't have the patch, if someone wants to ask me that :)

Code: Select all

1. Patch SG0005086 Release Note

This release note describes patch SG0005086 to IRIX
6.5/6.5SE and 6.5.1 to 6.5.20 inclusive.

This patch contains the IRIX 6.5.21 versions of the
installation tools inst, swmgr, showfiles and showprods.
Therefore, IRIX 6.5.21 itself will replace this patch.

If you are using the "live installation" method to upgrade
from any 6.5-based release prior to 6.5.21 to another from
6.5.21 onwards, you must first install this patch to enable
inst to handle certain additions to the format of the
distribution images.  If you have 6.5.21 installed on your
system, the required changes are already contained in inst.
Alternatively, a "miniroot install" will allow you to
install versions from 6.5.21 onwards without any patching.
You should be sure to use the miniroot contained in the
distribution to which you are upgrading.

This patch is also useful to install on any 6.5-based
release to take advantage of bug fixes made to the
installation tools (see below.)

So this means that live installing 6.5.21 from pre 6.5.21 IRIX is not possible without this patch. Hmmm, makes you wonder about what happens next....
hmmm, i will attempt a build next week, when i'm at camp. Other builds i will attempt are:

Faad2 (We almost had this!)

If anyone can come up with some more packages to build, let me know...
Heh, you beat me to it! I have it compiled as well but the build process took ages. I actually have a mips3 build because of my Challenge S. If someone needs that instead of the mips4 build let me know.

Celestia linkage is a minefield, even in MPRO74. Next week I'll have another go at it.

Transcode is full wil gcc hardcoded flags. This will take quite some time

Lisa, you're right about terragen. No open source for that one...

wxPython looks very funky...

Xmame/Xmess 0.71.1 builds are done. I will attempt a tardist for these binaries complete with documentation and rc files, because lately we seem to get a lot of questions about them.

I was very busy reviving two SGI's an R3K Indigo and a R10K O200. Battling sleep, i have to remove the PROM password from the R3K indigo, and the PROM of the O200 was so old it wouldn't even netboot at 100 Mbit full duplex. Try finding 10 mbit devices in these lanparties these days :)

Both are running fine now and both owners think these machines are absolutely wonderful.

Now if y'all excuse me, i have to get my booze to the sitting corner and socialize :)
Just ported it to my Linux machine with Hardware openGL support (Matrox)

Very nice i must say! Didn't got to hear the sounds, but the game runs smoothly. I had to adjust the makefile as follows:

Code: Select all

# Makefile for CGV practicals
# UoY 2002

CC = gcc

INCLUDE = -I/usr/local/include `libmikmod-config --cflags`
LINKS = -L/usr/local/lib/ -L/usr/X11R6/lib -lglut -lGL -lGLU -lX11 -lXi -lXmu -lm  `libmikmod-config --libs` -lpthread


%: %.c
$(CC) $(INCLUDE) $(CFLAGS) -o [email protected] $< $(LINKS) $(LDFLAGS)

all: encrusted

I had to install libglut-3.7 but that wasn't hard.
Oh, and i replaced 'fhypot' with 'hypot' to make it compile.
Welcome Martijn (yet another SGI Dutchie) :)

Thanks for the info. I'll update this in the announcement page
directedition wrote: With the way things are going over here, I just might learn Dutch and join those ranks. The Netherlands are a sort of Mecha for hackers.

Why won't America just annexe Dutchieland, so "All your 31337 H4X0r5 are belong to us"? :)

Save the coffeeshops though, it has an almost religious status over here (not in the last place, that you can buy Reefer in such a shop)

If in dire need for Reefer, the sentence: "Ik mot effe blowen, waar ken ik hash score ?" should suffice...
LoWeN wrote: I am agree with semy-fly that should figure out on the main page with a big red warning...

I've done that, but maybe Neko can help me pump the layout a bit bigger :)

Let's hope this information reaches as many people as possible...
chicago-joe wrote: The only thing I have seen so far, is a warning at startup about the "fast clock" not being able to run faster than 1000Hz :?
I'm not sure what this means, but I will look into it.

A strings from the 6.5.21 O2 firmware:

Code: Select all

neo:~/prom/ip32prom>strings ip32prom.image | grep -i fast
Ph_LogFast20Map: map 0x%x reg_value 0x%x
neo:~/prom/ip32prom>strings ip32prom.image | grep -i clock
clock: /2
clock: /3
clock: /4
Warning: time invalid, resetting clock to epoch.
Initialized tod clock.

So nothing in the PROM, must be IRIX then.

BTW a grep on '5200' displays nothing. Apparently by the firmware, a RM5200 is identical in PROM functionality compared to the R5000.
Thanks for the nice tardists! I've put a blog entry on the main page