History

19/05/2004 Callisto dies :-(
01/02/2001 Upgraded to 2.4.1.
25/01/2001 Upgraded to 2.4.1-pre10.
15/01/2001 Upgraded to 2.4.1-pre2.
05/01/2001 Upgraded to 2.4.0.
04/01/2001 Upgraded to 2.4.0-prerelease-ac5.
21/12/2000 Upgraded to 2.4.0-test13-pre3.
20/12/2000 Finished repartitioning. My system had become one complex living system of soft links.
18/12/2000 Finally my firewall is stable again, with a `new' disk.
28/11/2000 The disk in my firewall (2.5" SCSI disk in a DEC UDB) died :-(
24/11/2000 Upgraded to 2.4.0-test11.
20/6/2000 Upgraded to 2.4.0-test1-ac10.
5/6/2000 Upgraded to 2.4.0-test1-ac7.
4/6/2000 Upgraded to 2.4.0-test1.
21/3/2000 Upgraded to 2.3.99-pre3.
14/3/2000 Upgraded to 2.3.51.
7/3/2000 Upgraded to 2.3.50.
1/3/2000 Upgraded to 2.3.48.
24/2/2000 Upgraded to 2.3.47.
8/2/2000 Upgraded to 2.3.42.
28/1/2000 Upgraded to 2.3.41.
22/12/1999 Upgraded to 2.3.22.
05/10/1999 Upgraded to 2.3.18.
26/07/1999 Upgraded to 2.3.10.
14/07/1999 The MESH driver works again in FAST-10 mode with my Fireball, but the solution is not ideal.
15/06/1999 Upgraded to 2.3.6.
08/06/1999 Upgraded to 2.3.5. Or: back in sync again!
07/06/1999 Upgraded to 2.3.3.
Upgraded to 2.3.4.
31/05/1999 Since booting from floppy is very unreliable (OF always fails reading sufficient data from my floppies), I moved the Fireball to the MESH chain again, so I could use that disk for booting. I discovered that the MESH driver in 2.2.7 can no longer cope with FAST-5/10 devices. Since the Conner did async only, I never noticed. Hence back to 2.1.130, but at least callisto is alive again!
30/05/1999 The old Conner boot disk died. Although Amanda takes daily backups, I got my first broken DDS tape two days earlier, as Murphy suggests :-(
05/05/1999 Upgraded to 2.2.7.
04/05/1999 Upgraded to 2.2.6.
27/04/1999 Wrote an i2c driver for the Hydra Mac I/O. Now I can read the contents of the SPD EEPROMs on my SDRAM DIMMs.
22/03/1999 Upgraded to 2.2.4pre6 (finally).
26/01/1999 Upgraded to 2.2.0.
21/01/1999 Upgraded to 2.2.0-pre9 aka 2.2.0-final.
19/01/1999 Upgraded to 2.2.0-pre8.
14/01/1999 Upgraded to 2.2.0-pre7.
12/01/1999 Upgraded to 2.2.0-pre6.
05/01/1999 Upgraded to 2.2.0-pre4.
29/12/1998 Upgraded to 2.2.0-pre1.
28/12/1998 Got an old Conner 40 MB SCSI disk, to be used as boot disk. Since my Open Firmware can't boot from the Sym875, I need at least one device on the MESH chain to boot from. Old thrown-away 40 MB disks are perfect for that :-)

Moved all other SCSI devices to the Sym875 chain. It's a real shame the U-SCSI Fireball still does only 3.5 MB/s. If you ever wondered why a Viking is much more expensive, now you know why.

15/12/1998 Installed a DC-2976UW SCSI adapter. Now my Viking II flies at 17 MB/s. Connected to the MESH, it did 6 MB/s only. The Sym875 also causes less interrupt stress: a simple read benchmark with hdparm generates only 256 interrupts, compared to approx. 3500 with the MESH.
08/12/1998 My CHRP box now serves my `brand new' Tektronix XP21 X-terminal.

Hmm, since the XP21 has a full MC68030, it could run Linux/m68k (in theory). Anyone with technical specs? :-)

04/12/1998 Integrated the acceleration patches for XF68_FBDev in the XFree86 3.3.3 tree, including some more patches from the accelerated Xpmac, which is maintained by Ruyichi Oikawa.
02/12/1998 Upgraded to 2.1.130.
20/11/1998 Upgraded to 2.1.129.
15/11/1998 Upgraded to 2.1.128.
09/11/1998 Upgraded to 2.1.127.
02/11/1998 The first XF68_FBDev with ATI Mach64 hardware acceleration is serving my display.
27/10/1998 Upgraded to 2.1.126.
11/10/1998 Upgraded to 2.1.125.
14/09/1998 Upgraded to 2.1.121.
07/09/1998 Upgraded to 2.1.120.
01/09/1998 Upgraded to 2.1.119.
27/08/1998 Upgraded to 2.1.118.
19/08/1998 Added the section about 2-bank SDRAM.
13/08/1998 Upgraded to 128 MB RAM, and 7.7 GB disk.

Finally got the correct memory and an UW68-U50 SCSI convertor.

10/08/1998 Upgraded to 2.1.115.
04/08/1998 Upgraded to 2.1.114. Sound and modules are working fine now.
30/07/1998 Upgraded to 2.1.112.
29/07/1998 Trying to get modules to work.
26/07/1998 Upgraded to 2.1.111.
19/07/1998 Upgraded to 2.1.109.
14/07/1998 Sent back the 4-bank SDRAM. Back to 64 MB only :-(
08/07/1998 Released the first beta of my revamped ATI frame buffer device.
02/07/1998 Upgraded to 2x 128 MB SDRAM. Unfortunately this seem to be the incorrect type of SDRAM (4-bank instead of 2-bank), so I can use only half of the amount of memory :-( Still better than the old 64 MB, though. Well, I'll see whether I'll return them... Also received my new disk (4.5 GB Quantum Viking II). Unfortunately they delivered the wide version while I have narrow SCSI only. The shop promised to send me a free conversion cable... Let's wait and see...
29/06/1998 Upgraded to 2.1.107.
25/06/1998 Linus released 2.1.107, including the Abstract Console Driver patches.
17/06/1998 Sent the full Abstract Console Driver patches to Linus. Let's hope he accepts them.
16/06/1998 Upgraded to 2.1.106.
11/06/1998 Upgraded to 2.1.105.
21/05/1998 Finally hardware accelerated text mode (using bitblt) works on my ATI.
Upgraded to 2.1.103.
18/05/1998 Upgraded to 2.1.102.
17/05/1998 Upgraded to 2.1.101.
05/05/1998 Finally I found the LongTrail specs on ftp.austin.ibm.com
Upgraded to 2.1.99.
28/04/1998 Upgraded to 2.1.98. Kernels compiled with egcs-2.90.25 seem to be OK now.
24/04/1998 Upgraded to 2.1.97
13/04/1998 Upgraded to 2.1.95
31/03/1998 Finally I got the PS/2 keyboard working under Linux. Open Firmware incorrectly initialized the interrupt type for the PS/2 keyboard and mouse controller in the Super I/O chip. On newer boards, OF does it right.
30/03/1998 Upgraded to 2.1.91
11/03/1998 Upgraded to 2.1.89
12/02/1998 Upgraded to 2.1.86
10/02/1998 Upgraded to 2.1.85
04/02/1998 Finally the floppy driver works, thanks to a patch from Michael Wachs. Cause: code that was not endianness-clean (again).
28/02/1998 Upgraded to 2.1.80
27/02/1998 Got virtual screens, panning and dynamic mode switching working under X using the ATI frame buffer device (atyfb). Panning also greatly speeds up text console scrolling.
21/02/1998 Upgraded to 2.1.79
I should tell a lot about what happened in between:
  • Frame buffer device stuff for ATI video boards, cfr. the unified console driver.
  • IDE (Peter De Schrijver)
  • ADB busmouse emulation (Michael Schmitz et al.)
  • XF68_FBDev
  • ...
23/12/1997 Upgraded to 2.1.72.
07/12/1997 Upgraded to 2.1.70.
30/11/1997 /proc/cpuinfo shows the memory and cache configuration (thanks to the VAS96011/12 `Golden Gate II' chipset docs I've got from VLSI Technology).
24/11/1997 Upgraded to 2.1.65.
23/11/1997 /proc/cpuinfo shows the CHRP model.
13/11/1997 Upgraded to 2.1.63.
11/11/1997 Installed Netscape 4.04.
06/11/1997 Upgraded to 2.1.62.
03/11/1997 Upgraded to 2.1.61.
30/10/1997 Upgraded to 2.1.60.
29/10/1997 Got an ATI Xclaim 3D (4 MB SGRAM). Console and X work now.
28/10/1997 Added lmbench results.
21/10/1997 Upgraded to 2.1.59.
15/10/1997 Merged in Paul Mackerras' patches: support for ADB, OF device tree interrogation and /proc/device-tree.
14/10/1997 Borrowed an ADB keyboard and mouse.
9/10/1997 Upgraded to 2.1.57.
24/09/1997 Upgraded to 2.1.56.
17/09/1997 Added a 32 MB SDRAM DIMM. The total memory size is now 64 MB.
17/09/1997 Upgraded to 2.1.55. The boot loader now uses OF instead of hardware poking.
15/09/1997 Fixed a very annoying problem (messed up serial output during bootup) in the serial code. First half-working attempts for a bootloader that uses Open Firmware calls instead of hardware poking (based on the PowerMac coffboot code).
06/09/1997 Upgraded to Linux 2.1.53 and 2.1.54 and applied Miquel van Smoorenburg's serial console patches. Multi user mode works and I get a login prompt on my terminal. Shutdown reboots the machine.
30/08/1997 Booting kernels from a msdos partition on disk (scsi/disk@6,0:1,zimage) works. The machine can compile its own kernels: it takes about 9 minutes, versus 14 for crosscompilation on a 133 MHz Pentium.
25/08/1997 I extracted RedHat/PPC archives from DAT to hard disk. I can boot from hard disk (still using a kernel on a floppy) into single user mode. Mounting CD-ROMs works, running X programs and displaying them on a Linux/ix86 box works, too.
24/08/1997 Upgraded to Linux/PPC 2.1.51. The MESH SCSI driver detects all my SCSI devices.
17/08/1997 OpenPIC interrupts work fine now. 8259 interrupts mainly work, except for some problems with the keyboard and the serial port. I ported the MESH SCSI driver for the PowerMac to my machine, but it has some problems with the SCSI device detection. I switched from the Tulip Ethernet driver to de4x5 and managed to mount a root file system using NFS root.
03/08/1997 Added support for OpenPIC. OpenPIC interrupts seem to work, but legacy 8259 interrupts still don't come through.
19/07/1997 I made the timer work by using the decrementer (cfr. Linux/pmac). Result: 399.77 BogoMIPS. The kernel boots until the moment to mount the root filesystem. Interrupts still don't come through.
12/07/1997 Upgraded to Linux/PPC 2.0.30: now the Linux boot code finds out the exact amount of memory by querying Open Firmware. Still lock up while calibrating the delay loop, as expected.
11/07/1997 The kernel starts, PCI devices are identified, and it hangs while calibrating the delay loop because I don't know yet how to enable interrupts.
06/07/1997 The Linux/PPC 2.0.28 kernel is decompressed correctly, but the kernel crashes while initializing the MMU tree
05/07/1997 ISA I/O space is located at 0xf8000000, and programming the serial port works.
28/06/1997 I can load and execute very simple ELF programs under Open Firmware.
21/06/1997 The Open Firmware monitor works; Open Firmware finds all my SCSI devices.
13/06/1997 Arrival of the support hardware; assembly of the most essential parts.
30/05/1997 Arrival of the Open Firmware Flash ROM.
16/05/1997 Arrival of the motherboard.


This page is maintained by Geert Uytterhoeven.
$Date: 2006-01-25 21:24:49 $