basilisk-devel — Basilisk II development-related discussion

Re: [B2-devel] No space left on CVS server

[Christian B. <Chr...@un...>]

Hi!

On Wed, May 14, 2003 at 08:56:43AM +0200, Gwenole Beauchesne wrote:
> I have just tried to commit a few things and I got
> Insufficient disk space; try again later

Hm, the maillog files have filled with time-out messages related to
SourceForge's mail server. I've cleared up some space now.

Bye,
Christian

-- 
  / Coding on PowerPC and proud of it
\/ http://www.uni-mainz.de/~bauec002/

[B2-devel] No space left on CVS server

[Gwenole B. <gb...@di...>]

Hi,

I have just tried to commit a few things and I got
Insufficient disk space; try again later
Insufficient disk space; try again later
returntosender: cannot select queue for downcvs
Insufficient disk space; try again later
returntosender: cannot select queue for postmaster
Error writing control file tfh4E6o63K006937: No space left on device

Could please have a look?

Thanks,
Gwenole

Re: [B2-devel] SheepShaver progress

[Gwenole B. <gb...@di...>]

>> Does someone know what is supposed to be lowmem global at 0x1ca0?
>
> The trap address for $aba8 "stColorTab". Some internal QuickDraw stuff,
> I suppose...

I ran into another problem with MacOS 8.6. The one from the iMac DV 
install CD doesn't work either in native or in emulated mode. But the 
emulator helps to get a trace. ;-)

So, that's a NewWorld ROM from that 8.6 CD. Sorry, I don't know about 
the version. It appears that the .Sony driver doesn't get replaced, 
likewise for the PCFloppy ndrv. This leads to the ".Sony" driver open() 
(in InstallDrivers) to do nothing. Henceforth, the handle in the UTable 
is dangling to NULL. And since there is some garbage at 0x0000, 
dereferencing that heads to a segfault.

In the following trace, "*" denotes that the instruction was not run 
but is simply here for reference wrt. the natural control flow.

40814ae2: 205f                  movea.l (a7)+,a0
40814ae4: 225f                  movea.l (a7)+,a1
40814ae6: 221f                  move.l  (a7)+,d1
40814ae8: 241f                  move.l  (a7)+,d2
40814aea: 245f                  movea.l (a7)+,a2
40814aec: 4a40                  tst.w   d0
40814aee: 584f                  addq.w  #4,a7
40814af0: 4e75                  rts

408450b2: a260                  HFSDispatch				*
408450b4: 0c40 ffd5             cmpi.w  #$ffd5,d0
408450b8: 6726                  beq.s   $408450e0
408450ba: 4a40                  tst.w   d0
408450bc: 661c                  bne.s   $408450da
408450be: 2a28 0030             move.l  ($0030,a0),d5
408450c2: 0828 0004 001e        btst    #$04,($001e,a0)
408450c8: 6650                  bne.s   $4084511a
408450ca: 70d0                  moveq   #$ffffffd0,d0	*
408450cc: 600c                  bra.s   $408450da		*

4084511a: b5fc 656d 7074        cmpa.l  #$656d7074,a2
40845120: 6710                  beq.s   $40845132
40845122: b5fc 7472 7368        cmpa.l  #$74727368,a2
40845128: 6708                  beq.s   $40845132
4084512a: b5fc 6465 736b        cmpa.l  #$6465736b,a2
40845130: 6618                  bne.s   $4084514a
40845132: 0807 000a             btst    #$0a,d7			*
40845136: 6712                  beq.s   $4084514a		*

4084514a: 486e ff86             pea     ($ffffff86,a6)
4084514e: 2f0a                  move.l  a2,-(a7)
40845150: 2f05                  move.l  d5,-(a7)
40845152: 4eba 011c             jsr     ($40845270,pc)

40845270: 2f30 81e2 2018 0030   move.l  
([VectorPtr+4632],$00000030),-(a7)
40845278: 4e75                  rts

408451b0: 4e56 ffe6             link.w  a6,#$ffe6
408451b4: 2f0b                  move.l  a3,-(a7)
408451b6: 2078 02b6             movea.l ExpandMem,a0
408451ba: 2028 008c             move.l  ($008c,a0),d0
408451be: 6700 00a4             beq     $40845264

40845264: 265f                  movea.l (a7)+,a3
40845266: 4e5e                  unlk    a6
40845268: 205f                  movea.l (a7)+,a0
4084526a: defc 000c             adda.w  #$000c,a7
4084526e: 4ed0                  jmp     (a0)

40845156: 0c85 ffff 0000        cmpi.l  #$ffff0000,d5
4084515c: 6504                  bcs.s   $40845162
4084515e: 3005                  move.w  d5,d0
40845160: 600e                  bra.s   $40845170
40845162: 206e 0010             movea.l ($0010,a6),a0
40845166: 3084                  move.w  d4,(a0)			BREAK
[follow-up]
40845168: 206e 000c             movea.l ($000c,a6),a0
4084516c: 2085                  move.l  d5,(a0)
4084516e: 7000                  moveq   #$00,d0
40845170: 3d40 001c             move.w  d0,($001c,a6)
40845174: 0c46 ffff             cmpi.w  #$ffff,d6
40845178: 6716                  beq.s   $40845190
4084517a: 3f06                  move.w  d6,-(a7)
4084517c: a99a                  CloseResFile
4084517e: 558f                  subq.l  #2,a7
40845180: a9af                  ResError
40845182: 301f                  move.w  (a7)+,d0
40845184: 670a                  beq.s   $40845190
40845186: 4a6e 001c             tst.w   ($001c,a6)
4084518a: 6604                  bne.s   $40845190
4084518c: 3d40 001c             move.w  d0,($001c,a6)
40845190: 3f2e fec6             move.w  ($fffffec6,a6),-(a7)
40845194: a998                  UseResFile

At "BREAK", we have a problem because we have the following data (a6 
contains 217f2b58):
217f2b58: 217f2b76 40844d76 00000000 217f2d24  '!.+v@.Mv....!.-$'
217f2b68: 00000000 00006578 746effff 2009217f  '......extn.. .!.'
so, we are actually writing garbage to 0x0000 which will lead to a 
crash later.

I believe this may be because the .Sony driver replacement is missing. 
Indeed, with an oldworld ROM that area holds a valid pointer but SS 
would crash later. If anyone has interesting lights, I am a taker. ;-)

Bye,
Gwenole.

Re: [B2-devel] SheepShaver progress

[Christian B. <Chr...@un...>]

Hi!

On Sun, May 04, 2003 at 01:14:26PM +0200, Gwenole Beauchesne wrote:
> >SheepShaver is finally starting to boot with the "old" CPU core 
> >derived from Microlib's.

Cool. :-)

> Does someone know what is supposed to be lowmem global at 0x1ca0?

The trap address for $aba8 "stColorTab". Some internal QuickDraw stuff,
I suppose...

Bye,
Christian

-- 
  / Coding on PowerPC and proud of it
\/ http://www.uni-mainz.de/~bauec002/

[B2-devel] Re: basilisk-devel digest, Vol 1 #215 - 3 msgs

[Nigel P. <ni...@in...>]

> I believe a interesting feature of 1.0 is also the ability to run B2
> under MacOS X. I don't know the current status though.

Summary: OS X's memory addressing prevents the JIT core form being used.

Built under  Running under  Allows
-----------------------------------------------------------------
     10.1      10.1, 10.2    REAL_ADDRESSING=0 DIRECT_ADDRESSING=1
     10.2		      10.2    DIRECT_ADDRESSING=0 (*)
-----------------------------------------------------------------
(*) Setting DIRECT_ADDRESSING=1 causes the emulator
     to never write to its screen. I wish I knew why

--
| Nigel Pearson, ni...@in...   | "Now the world has gone to 
bed,
| Telstra BI&D, Sydney, Australia         |  Darkness won't engulf my 
head,
|   Office: 8255 4222    Fax:  8255 3153  |  I can see by infrared,
|   Mobile: 0408 664435  Home: 9792 6998  |  How I hate the night." 
-Marvin

Re: [B2-devel] New version?

[Gwenole B. <gb...@di...>]

Hi,

> I'm following this list for quite a while (although not many emails 
> here ;)
> and I was beginning to wonder - how about creating a new version with 
> the
> current CVS version?

Concerning the JIT compiler, it appears to be quite stable nowadays. 
The only bug left is related to the translation of FPU instructions in 
longish blocks. Otherwise, the "P4" related bug should be fixed in CVS. 
I had ideas to improve the JIT performance but it's indeed not critical 
for a release.

Concerning other parts, I got a report that fullscreen DGA broke again 
but I have not tried that recently.

I believe a interesting feature of 1.0 is also the ability to run B2 
under MacOS X. I don't know the current status though.

Bye,
Gwenole.

[B2-devel] New version?

[Hetz B. H. <he...@wi...>]

Hi,

I'm following this list for quite a while (although not many emails here ;) 
and I was beginning to wonder - how about creating a new version with the 
current CVS version?

According to the official web site: 
http://www.uni-mainz.de/~bauec002/B2Main.html - the official latest version 
is 0.9-1, dating May 2001, and I've seen some work done by Gwenole...

Is it possible? I'm sure that many people would be delighted to hear about a 
new version (if it helps - I'll publish it in slashdot - I'm a slashdot 
author)

Thanks,
Hetz

Re: [B2-devel] SheepShaver progress

[Gwenole B. <gb...@di...>]

Hi,

> SheepShaver is finally starting to boot with the "old" CPU core 
> derived from Microlib's. However, either it is slow, or this really 
> hangs after the boot logo has appeared and a few 68k A-Traps got run.

Beat me, I forgot to implement get_resource() trampoline. ;-)

Now, it is crashing at the same place as in native mode. That's because 
MacOS is trying to write to ROM. There is a sigsegv catcher and 
instruction skipper but only for PPC. I will try to operate likewise to 
what is done in B2 with SCRATCHMEM_SUBTERFUGE so that we don't need to 
write an instruction skipper on platforms using the ppc emulator.

Does someone know what is supposed to be lowmem global at 0x1ca0?

Bye,
Gwenole.

[B2-devel] SheepShaver progress

[Gwenole B. <gb...@di...>]

Hi,

SheepShaver is finally starting to boot with the "old" CPU core derived 
from Microlib's. However, either it is slow, or this really hangs after 
the boot logo has appeared and a few 68k A-Traps got run. Probably the 
UAE core should be used one day to replace 68k emulator...

Since SheepShaver is not little-endian aware yet, that experiment ran 
on PPC.

Christian, if you don't mind, I will probably move the ExecutePPC to 
some ExecuteNative and replace direct PPC functions invocations with 
"trampolines" serviced in the CPU emulator ala syscall. e.g. for 
VideoDoDriverIO et al. The current test approach I took is quite ugly 
though. Sources and patches are located at the usual places.

Bye,
Gwenole.

Re: [B2-devel] SheepShaver SIGBUS handler

[Gwenole B. <gb...@di...>]

> And EA is not aligned on 4-byte boundaries as stmw requests. I don't 
> know why it's so "recently". The workaround is to hook the 
> sigsegv_handler to SIGBUS and set ignoresegv to true in the prefs > file.

Hmm, reading it again, I should at least emulate unaligned writes and 
reads.

BTW, current CVS builds with gcc >= 3. I made EmulOp() extern "C". If 
this is a problem, I still can struggle with manual name mangling in 
asm_linux.S.

[B2-devel] SheepShaver SIGBUS handler

[Gwenole B. <gb...@di...>]

Hi,

I finally found out why SheepShaver no longer worked for me on earlier 
systems. It turns out kernel is finally generating bus errors on 
unaligned accesses (e.g. through stmw instructionts). Indeed, I got 
this while booting MacOS 8.1:

SIGSEGV
    pc 40a4f9dc     lr 40a4a0cc    ctr 40a4a078    msr 0000d032
   xer 00000000     cr 44000441
    r0 00000000     r1 20bf7eb0     r2 20186a68     r3 202d6c4e
    r4 20000a98     r5 40a4a0cc     r6 44000441     r7 201862e8
    r8 201869a8     r9 202d6970    r10 20186828    r11 20186c60
   r12 201868e8    r13 1008e380    r14 2000042c    r15 0000004b
   r16 2000042c    r17 20187570    r18 73637269    r19 408187ce
   r20 000007d2    r21 00000000    r22 00000816    r23 000007d2
   r24 202c7a60    r25 00000000    r26 20000428    r27 201fa5e0
   r28 202d69b8    r29 00000001    r30 202d6a3c    r31 202d6970
Welcome to the sheep factory.
40a4f9c4: 7ca802a6      mflr    r5
40a4f9c8: 7cc00026      mfcr    r6
40a4f9cc: 90a30000      stw     r5,$0000(r3)
40a4f9d0: 90c30004      stw     r6,$0004(r3)
40a4f9d4: 90230008      stw     r1,$0008(r3)
40a4f9d8: 9043000c      stw     r2,$000c(r3)
40a4f9dc: bda30014      stmw    r13,$0014(r3)
40a4f9e0: fc00048e      mffs    fr0
40a4f9e4: d9c30060      stfd    fr14,$0060(r3)
40a4f9e8: d9e30068      stfd    fr15,$0068(r3)
40a4f9ec: da030070      stfd    fr16,$0070(r3)
40a4f9f0: da230078      stfd    fr17,$0078(r3)
40a4f9f4: da430080      stfd    fr18,$0080(r3)
40a4f9f8: da630088      stfd    fr19,$0088(r3)
40a4f9fc: da830090      stfd    fr20,$0090(r3)

And EA is not aligned on 4-byte boundaries as stmw requests. I don't 
know why it's so "recently". The workaround is to hook the 
sigsegv_handler to SIGBUS and set ignoresegv to true in the prefs file. 
Is it OK with you? The same error probably prevented from running MacOS 
8.6 too.

Does someone know why r3 contents is unaligned?

Bye,
Gwenole.

Re: [B2-devel] CVS server down?

[Christian B. <Chr...@un...>]

Hi!

On Tue, Apr 08, 2003 at 11:08:37AM +1000, Nigel Pearson wrote:
> Or is the machine down or non routable?

Apparently, down.physik.uni-mainz.de is no longer reachable from outside
the campus. That's why I didn't notice any problems in the first place
(plus, I had email blockage).

As a temporary measure, I've aliased the machine to gromit.physik.uni-mainz.de
(134.93.128.231).

Bye,
Christian

-- 
  / Coding on PowerPC and proud of it
\/ http://www.uni-mainz.de/~bauec002/

[B2-devel] CVS server down?

[Nigel P. <ni...@in...>]

	For about the last 4 days, I have been unable to contact the machine
down.physik.uni-mainz.de for CVS access. Is the CVS repository moving?
Or is the machine down or non routable?

--
| Nigel Pearson, ni...@in...   | "Now the world has gone to 
bed,
| Telstra BI&D, Sydney, Australia         |  Darkness won't engulf my 
head,
|   Office: 8255 4222    Fax:  8255 3153  |  I can see by infrared,
|   Mobile: 0408 664435  Home: 9792 6998  |  How I hate the night." 
-Marvin

[B2-devel] CVS is down

[Gwenole B. <gb...@di...>]

Hi Christian,

I just wondered if you had a chance to look at the current CVS 
problems. Down seems to be down for now. ;-)

Thanks,
Gwenole.

[B2-devel] Patch to make ethertap module compile after making changes required in the README file

[Amartyo B. <ama...@ya...>]

When I tried to modify the ethertap module following
the instructions in the README file, I got the
following error message, which I have also posted on
the Sourceforge Basilisk Help forum, 
---------------------------------------------------------
gcc -D__KERNEL__ -I/usr/src/linux-2.4.19-16mdk/include
 -Wall -Wstrict-prototypes -Wno-trigraphs -O2
-fno-strict-aliasing -fno-common -fomit-frame-pointer
-pipe -mpreferred-stack-boundary=2 -march=i586
-DMODULE -DMODVERSIONS -include
/usr/src/linux-2.4.19-16mdk/include/linux/modversions.h
 -nostdinc -I
/usr/lib/gcc-lib/i586-mandrake-linux-gnu/3.2/include
-DKBUILD_BASENAME=ethertap  -c -o ethertap.o
ethertap.c
ethertap.c: In function `set_multicast_list':
ethertap.c:160: request for member `groups' in
something not a structure or union
make[2]: *** [ethertap.o] Error 1
make[2]: Leaving directory
`/usr/src/linux-2.4.19-16mdk/drivers/net'
make[1]: *** [_modsubdir_net] Error 2
make[1]: Leaving directory
`/usr/src/linux-2.4.19-16mdk/drivers'
make: *** [_mod_drivers] Error 2
-------------------------------------------------------
After looking at the source for ethertap.c and other
kernel source, I made some changes and now the
ethertap  module compiles when CONFIG_ETHERTAP_MC is
set.
Change to the directory above the kernel source
directory, eg '/usr/src' and apply the patch with
options -Np0
-------------------------------------------------------
diff -Naur linux-2.4.19-24mdk/drivers/net/ethertap.c
linux-2.4.19-24mdk-changed/drivers/net/ethertap.c
--- linux-2.4.19-24mdk/drivers/net/ethertap.c
2001-09-30 19:26:06.000000000 +0000
+++ linux-2.4.19-24mdk-changed/drivers/net/ethertap.c
2003-03-19 16:33:02.000000000 +0000
@@ -29,6 +29,7 @@
 #include <net/sock.h>
 #include <linux/netlink.h>
 
+#define CONFIG_ETHERTAP_MC 1
 /*
  *	Index to functions.
  */
@@ -101,7 +102,7 @@
 	ether_setup(dev);
 
 	dev->tx_queue_len = 0;
-	dev->flags|=IFF_NOARP;
+/*	dev->flags|=IFF_NOARP;*/
 	tap_map[dev->base_addr]=dev;
 
 	return 0;
@@ -156,7 +157,7 @@
 	}
 	lp->groups = groups;
 	if (lp->nl)
-		lp->nl->protinfo.af_netlink.groups = groups;
+		lp->nl->protinfo.af_netlink->groups = groups;
 }
 #endif
 
diff -Naur linux-2.4.19-24mdk/include/net/af_netlink.h
linux-2.4.19-24mdk-changed/include/net/af_netlink.h
--- linux-2.4.19-24mdk/include/net/af_netlink.h
1970-01-01 00:00:00.000000000 +0000
+++
linux-2.4.19-24mdk-changed/include/net/af_netlink.h
2003-03-19 16:33:02.000000000 +0000
@@ -0,0 +1,17 @@
+#ifndef _AF_NETLINK_H
+#define _AF_NETLINK_H
+struct netlink_opt
+{
+	u32			pid;
+ 	unsigned		groups;
+ 	u32			dst_pid;
+ 	unsigned		dst_groups;
+ 	unsigned long		state;
+ 	int			(*handler)(int unit, struct sk_buff *skb);
+ 	wait_queue_head_t	wait;
+ 	struct netlink_callback	*cb;
+ 	spinlock_t		cb_lock;
+ 	void			(*data_ready)(struct sock *sk, int bytes);
+};
+#endif
+
diff -Naur linux-2.4.19-24mdk/include/net/sock.h
linux-2.4.19-24mdk-changed/include/net/sock.h
--- linux-2.4.19-24mdk/include/net/sock.h	2003-01-30
20:14:18.000000000 +0000
+++ linux-2.4.19-24mdk-changed/include/net/sock.h
2003-03-19 16:33:02.000000000 +0000
@@ -106,6 +106,7 @@
 #include <asm/atomic.h>
 #include <net/dst.h>
 
+#include <net/af_netlink.h>
 
 /* The AF_UNIX specific socket options */
 struct unix_opt {
diff -Naur linux-2.4.19-24mdk/net/netlink/af_netlink.c
linux-2.4.19-24mdk-changed/net/netlink/af_netlink.c
--- linux-2.4.19-24mdk/net/netlink/af_netlink.c
2002-02-25 19:38:14.000000000 +0000
+++
linux-2.4.19-24mdk-changed/net/netlink/af_netlink.c
2003-03-19 16:33:02.000000000 +0000
@@ -41,6 +41,7 @@
 #include <linux/smp_lock.h>
 #include <net/sock.h>
 #include <net/scm.h>
+#include <net/af_netlink.h>
 
 #define Nprintk(a...)
 
@@ -48,19 +49,6 @@
 #define NL_EMULATE_DEV
 #endif
 
-struct netlink_opt
-{
-	u32			pid;
-	unsigned		groups;
-	u32			dst_pid;
-	unsigned		dst_groups;
-	unsigned long		state;
-	int			(*handler)(int unit, struct sk_buff *skb);
-	wait_queue_head_t	wait;
-	struct netlink_callback	*cb;
-	spinlock_t		cb_lock;
-	void			(*data_ready)(struct sock *sk, int bytes);
-};
 
 static struct sock *nl_table[MAX_LINKS];
 static DECLARE_WAIT_QUEUE_HEAD(nl_table_wait);

-------------------------------------------------------
Also, I read in the kernel documentation that the
ethertap module is obsolete and will go away soon and
to start using the tuntap module. Is Basilisk going to
be using the tuntap module in the future ?

__________________________________________________
Do you Yahoo!?
Yahoo! Platinum - Watch CBS' NCAA March Madness, live on your desktop!
http://platinum.yahoo.com

[B2-devel] New x86 code generator available

[Gwenole B. <gbe...@ma...>]

Hi,

I have finally integrated and fixed a new run-time assembler that is both
suitable for x86 and x86-64 code generation. The latter needs a few more
arrangements but it should now be much simpler to do the port there.

I hope I have not introduced any new regression. No performance increase
(nor decrease) is expected in current CVS sources. People willing to try
it out are invited to edit uae_cpu/compiler/codegen_x86.cpp and turn the
#if 0 to #if 1, just under the CLOBBER_ defines.

Note that I would like to obsolete the old run-time assembler. That's
simply because the new one is expected to be reused in several other
projects.

Bye,
Gwenole.

[B2-devel] JIT fixes pour Pentium 4 in CVS

[Gwenole B. <gbe...@ma...>]

Hi,

Some Pentium 4 happened to come next to me recently and I could fix
remaining problems with the JIT. So please, people if you had problems
with JIT on P4 (for example so garbled X video and such) please give it a
try.

Interestingly, one Speedometer 4 test showed performance around 110 times
faster than a Quadra 605. ;-) Though on average, that was only around 40
times.

Bye,
Gwenole

[B2-devel] Audio sample rate/format switching for Irix

[Brian J. <bjj...@us...>]

I've largely completed audio sample rate/format switching for Irix.
What I have (attached) mostly works.  I still need to hook up the
volume/mute controls, and perhaps the enter/exit stream functions.
But I have some questions:

- What is the sample format provided by MacOS for 8-bit stereo data?
  It doesn't appear to be packed bytes... I put in a unsigned->signed
  translation modeled on the one in the Amiga code (which, by the way,
  doesn't seem quite right, since it doesn't handle carries.  But that
  would only affect the low bits of the results, so perhaps it doesn't
  matter.)  That works fine for 8-bit mono, but 8-bit stereo seems to
  produce the requested sound (on one channel?) interleaved with
  garbage.  All the other combinations of 8/16-bit and mono/stereo
  work fine.

- Using 11KHz and 8KHz sound output (I only tried 16-bit stereo for
  those), most of the beep sounds in the MacOS 7.6 audio control panel
  don't work (they only produce silience), although "simple beep"
  works (it's a bit distorted) and the MacInTalk speech synthesizer
  does just fine.  Odd... any ideas?

- I noticed that stream_thread_cancel is never set to "false" in
  audio_oss_esd.cpp after it's set to "true" by close_audio().  So
  when changing the audio parameteres, won't stream_func() exit as
  soon as open_audio() restarts it?

Thanks,

						Brian

--------------------------------------------------------------------

  "ASCII stupid question, get a stupid ANSI!"
                                           -- Quoted by Khan Tran

Re: [B2-devel] Re: OS X CDROM problems

[Christian B. <cb...@th...>]

Hi!

On Wed, Jan 22, 2003 at 05:34:01PM +1100, Nigel Pearson wrote:
> 7.1 install disk on Intel Linux:
> ...
> CDROMOpen
>   DrvSts at 00003f08
>   TOC: 00140101 0014010000000200 0014AA0000090A41
>
> [...]
> Same CD on OS X:
> ...
> CDROMOpen
>  DrvSts at 00003e00
> ioctl(DKIOCCDREADTOC) read 20 bytes
>  TOC: 00120101 0014010000000200 0014aa0000090a41

TOC size is still 2 bytes off.

> 	That would seem to indicate that the data being read by Sys_read()
> is bad, but I find that very unlikely. Any thoughts to save me having
> to dump and compare 32KB of data?

That's what I was going to suggest next, actually... :-)

Bye,
Christian

-- 
  / Coding on PowerPC and proud of it
\/ http://www.uni-mainz.de/~bauec002/

[B2-devel] Re: OS X CDROM problems

[Nigel P. <ni...@in...>]

	Thanks Christian. That info was helpful. It is now closer.
After a little extra debug in read_toc(), and an extra flag in the=20
ioctl():


7.1 install disk on Intel Linux:
...
CDROMOpen
   DrvSts at 00003f08
   TOC: 00140101 0014010000000200 0014AA0000090A41
   Lead-Out M 9 S 10 F 65
   HFS partition found at 65536, 156270 blocks
   adding drive 2

Same CD on OS X:
...
CDROMOpen
  DrvSts at 00003e00
ioctl(DKIOCCDREADTOC) read 20 bytes
  TOC: 00120101 0014010000000200 0014aa0000090a41
  Lead-Out M 9 S 10 F 65
Looking for HFS partitions on CD-ROM...
  block 0, signature '=00=B7' (00ff)
  block 1, signature '=00=00' (0000)



	So, it has valid TOC info, but never finds a valid HFS =
signature.

	That would seem to indicate that the data being read by =
Sys_read()
is bad, but I find that very unlikely. Any thoughts to save me having
to dump and compare 32KB of data?

--
| Nigel Pearson, ni...@in...   | "Reality is that which,   |
| Telstra BI&D, Sydney, Australia         |  when you stop believing  |
|   Office: 8255 4222    Fax:  8255 3153  |  in it, doesn't go away." |
|   Mobile: 0408 664435  Home: 9792 6998  |  Philip K. Dick - 'Valis' |

Re: [B2-devel] OS X CDROM problems

[Christian B. <cb...@th...>]

Hi!

On Thu, Jan 16, 2003 at 10:59:38AM +1100, Nigel Pearson wrote:
> 7.1 install disk on Intel Linux:
> ...
> CDROMOpen
>  DrvSts at 00003f08
>  TOC: 00140101 00140100

4-byte header:
  0014: 20 bytes total TOC size (including header)
    01: First track = 1
    01: Last track = 1

followed by 8 bytes per track (the last track is always the lead-out,
number 0xAA)
  00: unused
  14: addr/ctrl
  01: track number
  00: unused
  xx: unused
  xx: track start M
  xx: track start S
  xx: track start F

(maybe read_toc() should dump the entire TOC for debugging...)

So for 1 data track + 1 lead-out -> 2*8 bytes + 4 bytes header = 20 bytes total

Bye,
Christian

-- 
  / Coding on PowerPC and proud of it
\/ http://www.uni-mainz.de/~bauec002/

[B2-devel] OS X CDROM problems

[Nigel P. <ni...@in...>]

	I am working on the CDROM parts of sys_unix.cpp for Mac OS X,
but haven't gotten them to work yet. Annoyingly, Apple has different
IOCTLs to FreeBSD and NetBSD.

	The strange thing is that the DEBUG results I am getting are =
similar,
but slightly different to, the Linux ones:

7.1 install disk on Intel Linux:
...
CDROMOpen
  DrvSts at 00003f08
  TOC: 00140101 00140100
  Lead-Out M 9 S 10 F 65
  HFS partition found at 65536, 156270 blocks
  adding drive 2

Same CD on OS X:
...
CDROMOpen
  DrvSts at 00003e00
ioctl(DKIOCCDREADTOC) read 20 bytes
  TOC: 00120101 00140100
  Lead-Out M 0 S 160 F 205
Looking for HFS partitions on CD-ROM...
  block 0, signature '=00=B7' (00ff)
  block 1, signature '=00=00' (0000)
...


	Note that the second byte of the TOC differs. Any ideas?

--
| Nigel Pearson, ni...@in...   | "People say I'm strange,=20
does it
| Telstra BI&D, Sydney, Australia         |        make me a stranger?
|   Office: 8255 4222    Fax:  8255 3153  |  My best friend was born
|   Mobile: 0408 664435  Home: 9792 6998  |         in a manger"    -DC=20=

Talk

Re: [B2-devel] Major problem solved

[Carsten N. <de...@sn...>]

On Sat, 14 Dec 2002, Gwenole Beauchesne wrote:
> Hi,
>=20
> > But I can't go to V1.0 because it crashes:
>=20
> Specify which snapshot please, as the error you report should be fixed=20
> in current CVS. And have you finally got a correct ROM extractor?

BasiliskII_src_15012002.tar.gz

- the latest I found.

complete URL:
http://iphcip1.physik.uni-mainz.de/~cbauer/BasiliskII_src_15012002.tar.gz


The ROM extractor is OK, one just needs to split the image.
I simply had to know the correct ROM size. 8-)
But now since I know them for my two Mac models ... no need to worry.

>=20
> Bye,
> Gwenole.

Thank you.
Kind regards,

=09Carsten
--=20
Certify your minority, support spyware!
Soon to come as TCPA and Palladium from Microsoft & Co.

Re: [B2-devel] Major problem solved

[Gwenole B. <gbe...@ma...>]

Hi,

> But I can't go to V1.0 because it crashes:

Specify which snapshot please, as the error you report should be fixed 
in current CVS. And have you finally got a correct ROM extractor?

Bye,
Gwenole.

[B2-devel] Major problem solved

[Carsten N. <de...@sn...>]

Hi all,

my `unsupported ROM' problem occured because I erroneously assumed the RO=
M's size is 1MB.
This was because my ROM extractor always reads 2MB and the IIci ROM image=
 file contains junk in
every 2nd 512KB block.

But V1.0 still crashes.
So I went back to V0.9 and Basilisk II boots fine.
But I can't change the error sound, can't use AppleTalk, and can't set th=
e timezone.
There are probably other problems which I haven't found yet.
I think this could be a PRAM problem.

But I can't go to V1.0 because it crashes:

Basilisk II V1.0 by Christian Bauer et al.
Reading ROM file...
Using /dev/dsp audio output
do_handle_screen_fault: unhandled address 0x446a4700 [IP=3D0x80a414e]
D0: 00000000 D1: fffcffff D2: fffffffc D3: 0000000f
D4: 0003fffc D5: 00000000 D6: 00000012 D7: 00000000
A0: 0000007c A1: 00000002 A2: 00006658 A3: 00006a6c
A4: 0000624c A5: 04291700 A6: 00000000 A7: 020009ae
USP=3D00000000 ISP=3D020009ae MSP=3D00000000 VBR=3D00000000
T=3D00 S=3D1 M=3D0 X=3D1 N=3D1 Z=3D0 V=3D0 C=3D0 IMASK=3D0
FP0: 0 FP1: 0 FP2: 0 FP3: 0
FP4: 0 FP5: 0 FP6: 0 FP7: 0
N=3D0 Z=3D0 I=3D0 NAN=3D0
0402e8e8: 2815 28c4 c081 c284 8287 MOVE.L (A5),D4
next PC: 0402e8ea

Thank you for your patience, sorry for the noise,

=09Carsten

--=20
Certify your minority, support spyware!
Soon to come as TCPA and Palladium from Microsoft & Co.