[OpenRISC] Simple linux with OR1200

[piotrcz]

Hi,

I am new and I recently have started using OR1200.
I have done simple project with UART and SRAM memory to write from processor
some data to TestBench via UART.
I am using simulator for this.

I would like to try boot simple Linux on OR1200. My hardware contains the
following peripherals:
- INTGEN
- UART
- SRAM memory

I have prepared linux binary file, however I see that some code for
exceptions is in wrong address space. Should start from 0x100 address but
starts from 0x800 address. When I created assembler from binary file
everything looks ok.

I would like to put this simple Linux directly to the SRAM memory but it is
not working.

I tried to use bootloader which is in SROM memory, but this is not working
also.

I am getting one of two problems:
- Processor hangs in die_if_kernel function
- bus error - there is reading transaction on data bus from very high
address

I do not know what I am doing wrong.

This is first time when I try to boot linux on SoC.

Has anyone some simple example how to configure and prepare Linux for very
simple hardware with UART?

Best Regards,
Piotr



--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978.html
Sent from the OpenRISC mailing list archive at Nabble.com.

[Stefan Kristiansson]

Post by piotrcz
I have prepared linux binary file, however I see that some code for
exceptions is in wrong address space. Should start from 0x100 address but
starts from 0x800 address. When I created assembler from binary file
everything looks ok.

This sounds weird, how did you build the kernel?

Stefan

[piotrcz]

Hi,

I generate in the same way which is shown in
http://www.rte.se/blog/blogg-modesty-corex/installing-linux/2.11
But maybe I am doing something wrong.
Linux kernel works with or1ksim, however, I see 2 strange things:
1. Stack dump
init started: BusyBox v1.19.0.g------------[ cut here ]------------
WARNING: CPU: 0 PID: 1 at kernel/irq/handle.c:146 0xc0044fcc()
irq 2 handler 0xc0156eb8 enabled interrupts
Modules linked in:
CPU: 0 PID: 1 Comm: init Tainted: G W
3.13.0-rc7-or1ksim-13847-g43e4944-dirty #12
Stack dump [0xc7019c28]:
sp + 00: 0xc7019c28
sp + 04: 0xc7019c74
sp + 08: 0x00000009
sp + 12: 0xc0044fcc
sp + 16: 0x00000092
sp + 20: 0xc025562f
sp + 24: 0xc7019ea0
sp + 28: 0xc000acc0
sp + 32: 0x00000000
sp + 36: 0x00000001
sp + 40: 0xc025562f
sp + 44: 0x00000092
sp + 48: 0xc0044fcc
sp + 52: 0xc7019c84
sp + 56: 0x00000001
sp + 60: 0x00000000
sp + 64: 0x00000000
sp + 68: 0x00000002
sp + 72: 0xc000ad74
sp + 76: 0xc0255643
sp + 80: 0xc7019c84
sp + 84: 0xc70aee70
sp + 88: 0xc0044fcc
sp + 92: 0x00000002
sp + 96: 0xc0156eb8
sp + 100: 0xc70703d4
sp + 104: 0xc028f40c
sp + 108: 0xc03fa070
sp + 112: 0x00000000
sp + 116: 0x000000c7
sp + 120: 0xc7019ea0
sp + 124: 0x00000000
sp + 128: 0x00000000
sp + 132: 0xc7019e38
sp + 136: 0xc0045014
sp + 140: 0xc70703d4
sp + 144: 0xc0047b20
sp + 148: 0x00000002
sp + 152: 0xc00445ac
sp + 156: 0x00000002
sp + 160: 0xc0004ca8
sp + 164: 0xc7064714
sp + 168: 0xc7002b00
sp + 172: 0x000000c7
sp + 176: 0x000000c7
sp + 180: 0xc0005580
sp + 184: 0x0000807f
sp + 188: 0xc7019df0
sp + 192: 0xc7064714
sp + 196: 0xc7002b00
sp + 200: 0x00000001
sp + 204: 0x00000000
sp + 208: 0x00000001
sp + 212: 0x00000000
sp + 216: 0x00000000
sp + 220: 0xc00b9428
sp + 224: 0xc7018000
sp + 228: 0x00000001
sp + 232: 0x00000000
sp + 236: 0x0000001f
sp + 240: 0xc7002b00
sp + 244: 0x00000005
sp + 248: 0x000000c7
sp + 252: 0x00000000
sp + 256: 0x000000c7
sp + 260: 0x00008079
sp + 264: 0x000000c7
sp + 268: 0x0000000f
sp + 272: 0xc7019ea0
sp + 276: 0x00000000
sp + 280: 0x00000000
sp + 284: 0xffffffff
sp + 288: 0x00000000
sp + 292: 0xfffffff9
sp + 296: 0xc7019e38
sp + 300: 0xc009e6b8
sp + 304: 0xc7019e28
sp + 308: 0x00000000
sp + 312: 0xc00bf718
sp + 316: 0xffffffff
sp + 320: 0x01ffffff
sp + 324: 0x00000068
sp + 328: 0x00000000
sp + 332: 0x00000000
sp + 336: 0x000000a8
sp + 340: 0xc70aa000
sp + 344: 0xc7019dac
sp + 348: 0xc0424ce0
sp + 352: 0x00000000
sp + 356: 0xffffffe5
sp + 360: 0xc70647c0
sp + 364: 0xfffffffd
sp + 368: 0x00000000
sp + 372: 0x00000000
sp + 376: 0xc0424ce0
sp + 380: 0x00000000
sp + 384: 0xffffffe5
sp + 388: 0xc7064714
sp + 392: 0xfffffffd
sp + 396: 0x00000000
sp + 400: 0xc00790b0
sp + 404: 0xfffffffd
sp + 408: 0x000ce19c
sp + 412: 0x00000000
sp + 416: 0xc70647c4
sp + 420: 0x00000000
sp + 424: 0xc7064714
sp + 428: 0xc7002b00
sp + 432: 0x000000c7
sp + 436: 0x000000c7
sp + 440: 0x000000c7
sp + 444: 0xc7019ea0
sp + 448: 0x00000000
sp + 452: 0xc00b9380
sp + 456: 0x00000002
sp + 460: 0x19a14780
sp + 464: 0x00000002
sp + 468: 0x19a14780
sp + 472: 0x00000000
sp + 476: 0x00000000
sp + 480: 0xc007b434
sp + 484: 0xc7019e88
sp + 488: 0x000ce19c
sp + 492: 0x00000000
sp + 496: 0xc70af4a0
sp + 500: 0x00002820
sp + 504: 0xc70af4a8
sp + 508: 0x00000001
sp + 512: 0x000000c7
sp + 516: 0x00000f39
sp + 520: 0x0013c137
sp + 524: 0x00000000
sp + 528: 0x00000000
sp + 532: 0x00001000
sp + 536: 0xc7019ecc
sp + 540: 0xc70af4a0
sp + 544: 0xc7019ecc
sp + 548: 0x0013c070
sp + 552: 0x00000000
sp + 556: 0x00000000
sp + 560: 0x0013d084
sp + 564: 0x00000001
sp + 568: 0x00000000
sp + 572: 0x00000000
sp + 576: 0xc009d5b8
sp + 580: 0xc70af4a0
sp + 584: 0x00000000
sp + 588: 0x00000000
sp + 592: 0x00000000
sp + 596: 0xc70012a0
sp + 600: 0x00000000
sp + 604: 0x00000000
sp + 608: 0x00000000
sp + 612: 0x000000c7
sp + 616: 0x00001000
sp + 620: 0x00000000
sp + 624: 0x00000000
sp + 628: 0x00000000
sp + 632: 0x0013c070
sp + 636: 0x00001000
sp + 640: 0x00001000
sp + 644: 0xc009e484
sp + 648: 0xfffffff7
sp + 652: 0xc70af4a0
sp + 656: 0x00001000
sp + 660: 0x0013c070
sp + 664: 0x00000000
sp + 668: 0x00000000
sp + 672: 0xc009e720
sp + 676: 0x00000000
sp + 680: 0x00000000
sp + 684: 0x0013c028
sp + 688: 0x00000000
sp + 692: 0x0013c028
sp + 696: 0x001334f8
sp + 700: 0x00000000
sp + 704: 0x7fa71e20
sp + 708: 0xc0005698
sp + 712: 0x0000847e
sp + 716: 0x7fa71dac
sp + 720: 0x0013c028
sp + 724: 0x00000003
sp + 728: 0x0013c070
sp + 732: 0x00001000
sp + 736: 0x7fa71e64
sp + 740: 0x00000000
sp + 744: 0x00000000
sp + 748: 0x00109df4
sp + 752: 0x00000000
sp + 756: 0x0000003f
sp + 760: 0x00000000
sp + 764: 0x000e7994
sp + 768: 0x0013d078
sp + 772: 0x0000807e
sp + 776: 0x7fa71f28
sp + 780: 0xfffffff9
sp + 784: 0x001334f8
sp + 788: 0x00008079
sp + 792: 0x001334f9
sp + 796: 0x0000000f
sp + 800: 0x00000004
sp + 804: 0x00000000
sp + 808: 0x7fa71f24
sp + 812: 0xffffffff
sp + 816: 0x00000000
sp + 820: 0xfffffff9
sp + 824: 0x00000000
sp + 828: 0x00008079
sp + 832: 0x00000000
sp + 836: 0x00000000
sp + 840: 0x000e8054
sp + 844: 0x0000003f
sp + 848: 0x00000000
sp + 852: 0x00000000
sp + 856: 0x00000000
sp + 860: 0x00000000
sp + 864: 0x00000000
sp + 868: 0x00000000
sp + 872: 0x00000000
sp + 876: 0x00000000
sp + 880: 0x00000000
sp + 884: 0x00000000
sp + 888: 0x00000000
sp + 892: 0x00000000
sp + 896: 0x00000000
sp + 900: 0x00000000
sp + 904: 0x00000000
sp + 908: 0x00000000
sp + 912: 0x00000000
sp + 916: 0x00000000
sp + 920: 0x00000000
sp + 924: 0x00000000
sp + 928: 0x00000000
sp + 932: 0x00000000
sp + 936: 0x00000000
sp + 940: 0x00000000
sp + 944: 0x00000000
sp + 948: 0x00000000
sp + 952: 0x00000000
sp + 956: 0x00000000
sp + 960: 0x00000000
sp + 964: 0x00000000
sp + 968: 0x00000000
sp + 972: 0x00000000
sp + 976: 0x00000000
sp + 980: 0x00000000

[<c0044fcc>]
[<c000acc0>]
[<c0044fcc>]
[<c000ad74>]
[<c0044fcc>]
[<c0156eb8>]
[<c0045014>]
[<c0047b20>]
[<c00445ac>]
[<c0004ca8>]
[<c0005580>]
[<c00b9428>]
[<c009e6b8>]
[<c00bf718>]
[<c00790b0>]
[<c00b9380>]
[<c007b434>]
[<c009d5b8>]
[<c009e484>]
[<c009e720>]
[<c0005698>]
=======================
---[ end trace b64f93fd77ce5950 ]---

2. Please press Enter to activate this console. ifconfig: SIOCSIFADDR: No
such device

I am not suer if this is ok. Maybe this causes issues in simulation with
hardware

I have 3 files:
- vmlinux - binary file
- vmlinus.vmem - file for RTL simulation
- vmlinux.asm - asembler

I see something strange in these files:
- vmlinux
beginning of file contains some data and next some data are from address
00002100 (I think this is part for Reset exception)
- vmlinux.vmem
@00000000 7f454c46 01020100 00000000 00000000
@00000004 0002005c 00000001 c0000000 00000034
@00000008 0043e63c 00000000 00340020 00020028
@0000000c 001c0019 00000001 00002000 c0000000
@00000010 00000000 002791a0 0027a000 00000007
@00000014 00002000 00000001 0027c000 c027a000
@00000018 0027a000 0017f59c 00193ed8 00000007
@0000001c 00002000 00000000 00000000 00000000
... zeros
@00000840 19e0c029 a9ef8000 19a04000 e1ad7800
@00000844 44006800 15000000 00000000 00000000
@00000848 00000000 00000000 00000000 00000000
@0000084c 00000000 00000000 00000000 00000000
@00000850 00000000 00000000 00000000 00000000
@00000854 00000000 00000000 00000000 00000000
@00000858 00000000 00000000 00000000 00000000
@0000085c 00000000 00000000 00000000 00000000
@00000860 00000000 00000000 00000000 00000000
@00000864 00000000 00000000 00000000 00000000
@00000868 00000000 00000000 00000000 00000000
@0000086c 00000000 00000000 00000000 00000000
@00000870 00000000 00000000 00000000 00000000
@00000874 00000000 00000000 00000000 00000000
@00000878 00000000 00000000 00000000 00000000
@0000087c 00000000 00000000 00000000 00000000
@00000880 d400f078 b7c00040 a7de0001 bc1e0000
@00000884 d400507c 0c00000a d4000880 1820c028
@00000888 a8218b30 1bc04000 e3de0800 855e0000
@0000088c 1bc04000 e3de5000 843e0038 9c21fef0
@00000890 1bc04000 e3de0800 d41e6030 b5800020
@00000894 d41e6080 b5800040 d41e6000 85800078
@00000898 d41e6078 8580007c d41e6028 85800080
@0000089c d41e6004 d41e2010 b4800030 e1800004
@000008a0 abc00079 c000f040 1bc0c000 abde5078
@000008a4 c000f020 24000000 00000000 00000000
@000008a8 00000000 00000000 00000000 00000000

etc.

As you can see Reset exception code is from address 840 not from 100 address

- vmlinux.asm

./vmlinux: file format elf32-or32


Disassembly of section .text:

c0000000 <_stext>:
...
c0000100: 19 e0 c0 29 l.movhi r15,0xc029
c0000104: a9 ef 80 00 l.ori r15,r15,0x8000
c0000108: 19 a0 40 00 l.movhi r13,0x4000
c000010c: e1 ad 78 00 l.add r13,r13,r15
c0000110: 44 00 68 00 l.jr r13
c0000114: 15 00 00 00 l.nop 0x0
...

c0000200 <_dispatch_bus_fault>:
c0000200: d4 00 f0 78 l.sw 0x78(r0),r30
c0000204: b7 c0 00 40 l.mfspr r30,r0,0x40

I am not sure if these files are ok.

Can I use this vmem file in simulation to put directly to SRAM memory or
should I use some bootloader.

Thank you for any help.
Best Regards,
Piotr








--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978p4641981.html
Sent from the OpenRISC mailing list archive at Nabble.com.

[Stefan Kristiansson]

Post by piotrcz
- vmlinux - binary file

This is most likely not a binary file, but an elf.

Post by piotrcz
- vmlinus.vmem - file for RTL simulation
- vmlinux.asm - asembler
- vmlinux
beginning of file contains some data and next some data are from address
00002100 (I think this is part for Reset exception)
- vmlinux.vmem
@00000000 7f454c46 01020100 00000000 00000000
@00000004 0002005c 00000001 c0000000 00000034
@00000008 0043e63c 00000000 00340020 00020028
@0000000c 001c0019 00000001 00002000 c0000000
@00000010 00000000 002791a0 0027a000 00000007
@00000014 00002000 00000001 0027c000 c027a000
@00000018 0027a000 0017f59c 00193ed8 00000007
@0000001c 00002000 00000000 00000000 00000000
... zeros
@00000840 19e0c029 a9ef8000 19a04000 e1ad7800
@00000844 44006800 15000000 00000000 00000000
@00000848 00000000 00000000 00000000 00000000
@0000084c 00000000 00000000 00000000 00000000
@00000850 00000000 00000000 00000000 00000000
@00000854 00000000 00000000 00000000 00000000
@00000858 00000000 00000000 00000000 00000000
@0000085c 00000000 00000000 00000000 00000000
@00000860 00000000 00000000 00000000 00000000
@00000864 00000000 00000000 00000000 00000000
@00000868 00000000 00000000 00000000 00000000
@0000086c 00000000 00000000 00000000 00000000
@00000870 00000000 00000000 00000000 00000000
@00000874 00000000 00000000 00000000 00000000
@00000878 00000000 00000000 00000000 00000000
@0000087c 00000000 00000000 00000000 00000000
@00000880 d400f078 b7c00040 a7de0001 bc1e0000
@00000884 d400507c 0c00000a d4000880 1820c028
@00000888 a8218b30 1bc04000 e3de0800 855e0000
@0000088c 1bc04000 e3de5000 843e0038 9c21fef0
@00000890 1bc04000 e3de0800 d41e6030 b5800020
@00000894 d41e6080 b5800040 d41e6000 85800078
@00000898 d41e6078 8580007c d41e6028 85800080
@0000089c d41e6004 d41e2010 b4800030 e1800004
@000008a0 abc00079 c000f040 1bc0c000 abde5078
@000008a4 c000f020 24000000 00000000 00000000
@000008a8 00000000 00000000 00000000 00000000

This looks like it's generated from an elf, not a binary file.

You can use objcopy to generate the binary
or1k-elf-objcopy -O binary vmlinux vmlinux.bin

Stefan

[piotrcz]

silly mistake.

However, I have generate binary file and vmem file. Now exceptions are in
correct place but processor hangs.

The last commands read from memory are:
84 c4 00 00
d7 e1 4f fc
a4 c6 00 01

and this commands are from die_if_kernel
c0004734 <die_if_kernel>:
c0004734: 84 c4 00 00 l.lwz r6,0x0(r4)
c0004738: d7 e1 4f fc l.sw 0xfffffffc(r1),r9
c000473c: a4 c6 00 01 l.andi r6,r6,0x1
c0004740: bc 06 00 00 l.sfeqi r6,0x0
c0004744: 0c 00 00 06 l.bnf c000475c <die_if_kernel+0x28>
c0004748: 9c 21 ff fc l.addi r1,r1,0xfffffffc
c000474c: 9c 21 00 04 l.addi r1,r1,0x4
c0004750: 85 21 ff fc l.lwz r9,0xfffffffc(r1)
c0004754: 44 00 48 00 l.jr r9
c0004758: 15 00 00 00 l.nop 0x0
c000475c: 9c 21 00 04 l.addi r1,r1,0x4
c0004760: 85 21 ff fc l.lwz r9,0xfffffffc(r1)
c0004764: 03 ff ff d8 l.j c00046c4 <die>
c0004768: 15 00 00 00 l.nop 0x0

Piotr



--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978p4641983.html
Sent from the OpenRISC mailing list archive at Nabble.com.

[Geert Uytterhoeven]

Hi Piotr,

Post by piotrcz
However, I have generate binary file and vmem file. Now exceptions are in
correct place but processor hangs.
84 c4 00 00
d7 e1 4f fc
a4 c6 00 01
and this commands are from die_if_kernel
c0004734: 84 c4 00 00 l.lwz r6,0x0(r4)
c0004738: d7 e1 4f fc l.sw 0xfffffffc(r1),r9
c000473c: a4 c6 00 01 l.andi r6,r6,0x1
c0004740: bc 06 00 00 l.sfeqi r6,0x0
c0004744: 0c 00 00 06 l.bnf c000475c <die_if_kernel+0x28>
c0004748: 9c 21 ff fc l.addi r1,r1,0xfffffffc
c000474c: 9c 21 00 04 l.addi r1,r1,0x4
c0004750: 85 21 ff fc l.lwz r9,0xfffffffc(r1)
c0004754: 44 00 48 00 l.jr r9
c0004758: 15 00 00 00 l.nop 0x0
c000475c: 9c 21 00 04 l.addi r1,r1,0x4
c0004760: 85 21 ff fc l.lwz r9,0xfffffffc(r1)
c0004764: 03 ff ff d8 l.j c00046c4 <die>
c0004768: 15 00 00 00 l.nop 0x0

Do you start the kernel with r3 = 0?
If r3 contains random data, the kernel will assume it's a pointer to the FDT,
and may crash.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ***@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

[piotrcz]

I have added simple code to ROM before executing code from RAM:
0 : wb_dat_o <= 32'h18000000;
1 : wb_dat_o <= 32'h18600000;
2 : wb_dat_o <= 32'h18800000;
3 : wb_dat_o <= 32'ha8840100;
4 : wb_dat_o <= 32'h44002000;
5 : wb_dat_o <= 32'h15000000;

But result is the same.

Piotr



--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978p4641986.html
Sent from the OpenRISC mailing list archive at Nabble.com.

[Stefan Kristiansson]

Post by piotrcz
silly mistake.
However, I have generate binary file and vmem file. Now exceptions are in
correct place but processor hangs.
84 c4 00 00
d7 e1 4f fc
a4 c6 00 01
and this commands are from die_if_kernel
c0004734: 84 c4 00 00 l.lwz r6,0x0(r4)
c0004738: d7 e1 4f fc l.sw 0xfffffffc(r1),r9
c000473c: a4 c6 00 01 l.andi r6,r6,0x1
c0004740: bc 06 00 00 l.sfeqi r6,0x0
c0004744: 0c 00 00 06 l.bnf c000475c <die_if_kernel+0x28>
c0004748: 9c 21 ff fc l.addi r1,r1,0xfffffffc
c000474c: 9c 21 00 04 l.addi r1,r1,0x4
c0004750: 85 21 ff fc l.lwz r9,0xfffffffc(r1)
c0004754: 44 00 48 00 l.jr r9
c0004758: 15 00 00 00 l.nop 0x0
c000475c: 9c 21 00 04 l.addi r1,r1,0x4
c0004760: 85 21 ff fc l.lwz r9,0xfffffffc(r1)
c0004764: 03 ff ff d8 l.j c00046c4 <die>
c0004768: 15 00 00 00 l.nop 0x0

You'll need to back-track this a bit further, what was it doing before
die_if_kernel was called?

Stefan

[piotrcz]

Hi,

I have traced this in my project and I think I found what is wrong.
CPU is going to execute code for bus error (0x200 address) I checked that
before this there is transaction from some slave.
Address of this slave is (A86321B0), but I do not have any device which uses
this address. This causes watchdog error and bus error.

Can anyone help me how to solve this issue.
Probably some option in Linux configuration must be disabled but I do not
know which one.

Thank you for help.

Best Regards,
Piotr

PS. I hope that this is the root of my issue :)



--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978p4641993.html
Sent from the OpenRISC mailing list archive at Nabble.com.

[piotrcz]

Hi,

Has anyone idea what is wrong?
I think that I did not disable some device in Linux configuration, however I
do not know which one uses A86321B0 address.

Best Regards,
Piotr



--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978p4641994.html
Sent from the OpenRISC mailing list archive at Nabble.com.

[Geert Uytterhoeven]

Hi Piotr,

Post by piotrcz
Has anyone idea what is wrong?
I think that I did not disable some device in Linux configuration, however I
do not know which one uses A86321B0 address.

Can you please try wiring up _early_uart_init() and _emergency_print() inside
printk(), so we'll get a backtrace?

BTW, as you didn't reply to my question "Do you start the kernel with r3 = 0?",
A86321B0 could still be a random number left in a register...

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ***@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

[piotrcz]

Hi Geert,

I have put this answer:
"I have added simple code to ROM before executing code from RAM:
0 : wb_dat_o <= 32'h18000000;
1 : wb_dat_o <= 32'h18600000;
2 : wb_dat_o <= 32'h18800000;
3 : wb_dat_o <= 32'ha8840100;
4 : wb_dat_o <= 32'h44002000;
5 : wb_dat_o <= 32'h15000000;

But result is the same.

Piotr "

I added line 2 : wb_dat_o <= 32'h18800000; for ROM code to clear register
R3. Is it ok?

I will try to get some backtrace.

I think that this is probably some silly mistake of new user.

Best Regards,
Piotr



--
View this message in context: http://openrisc.2316802.n4.nabble.com/Simple-linux-with-OR1200-tp4641978p4641996.html
Sent from the OpenRISC mailing list archive at Nabble.com.