Amithlon Setup Guide

[SnkBitten]

Kernel compiling "starter" guide is available. If you have any troubles let me know, I haven't tested some of the changes I made in the setup (copying the .config file and cleaning up then copying the .config file back in place and doing 'make oldconfig' to NOT have to walk through the manual process of choosing all the configuration settings for the kernel build).

Anyway, it's on my site http://amithlon.snkbitten.com/ as a published Google Doc so it may change as I get feedback or find errors. This should be an exact pull of Milan's kernel 4 with his .config file.

It's noted in the guide and I'll note it here as well. Pay a lot of attention to the .config file. It's the options for the kernel and one add or one subtraction of an option can prevent the kernel from booting your Amithlon system.

[themamboman]

Just a couple of notes that I've found so far upon compiling the kernel for those who want to attempt this.

I'm using an early version of SnkBitten's guide and it's great! Just make sure to read it all the way through first.

First issue I encountered was getting the GPG key for the RPM updates. It kept failing trying to do it using the remote address. I had to use wget to get the key local, then run the rpm --import command on the local copy.

Also, the .config file can be extracted from milanca's diff file, starting at line 402, copy out 924 lines to a new file, replace the "+" with an empty space using gedit or something similar and saving it to the file .config

The ./drivers/block/rd.c file at line 88 has a compile error. For some reason, it wouldn't pick up the #define for the block size from the config file, so I just replaced that with 4096 in that line.

The whole kernel compiled in less than 2 minutes, which was surprising.

I haven't had a chance to test it yet, but the kernel image built was in less than 2 minutes. Same for the modules. The final modules show up in
/lib/modules/2.4.37.9/kernel/drivers/

during make modules_install, there was dependency warning for RTL 8169 network drivers but the output file was created.

[SnkBitten]

Originally Posted by themamboman

First issue I encountered was getting the GPG key for the RPM updates. It kept failing trying to do it using the remote address. I had to use wget to get the key local, then run the rpm --import command on the local copy.

I had that issue while walking through the guide and I re-keyed in the line again and it took...so I thought I had a type-o or such....I didn't....but it didn't work, then it did so I left it.

copying the .config file and cleaning and then copying it back and running 'make oldconfig' gave me zero compile errors (tested while creating the guide).

Compile time is very short. Today's hardware makes compiling older kernels a quick process so you can compile, test, make changes, compile and test again with very little time consumed.

[SnkBitten]

Here's kernel 2.4.37.11 tree patched. Inside the .tar file you'll also find my diff patch file "amithlon.patch" which I used to patch kernel 2.4.37.11.

http://amithlon.snkbitten.com/SnkBit...4.37.11.tar.gz

And if you just want the diff file...
http://amithlon.snkbitten.com/SnkBit...n-patch.tar.gz

The "amithlon.patch" file was created by running a diff between the stock linux-2.4.37.9 tree and Milan's linux-2.4.37.9-amithlon tree, creating the "diff"erences file which can then be used to patch the stock tree. I applied it to linux-2.4.37.11 tree and was able to compile a kernel that boots my VB Amithlon.

[themamboman]

Thanks for that SnkBitten! I'll build it here today and try it out tonight. I'm going to turn on a few hardware options for my own machine though.

For those interested, this .11 version is the final 2.4. There are a few minor improvements over the .9 that Milanca's kernel 4 used. Assuming no Amithlon-specific changes went in, here's the changelog:

Linux Kernel-specific changes:

Summary of changes from v2.4.37.10 to v2.4.37.11
============================================
drivers/net/eql.c: prevent reading uninitialized stack memory
sys_semctl: fix kernel stack leakage
x25: Prevent crashing when parsing bad X.25 facilities
rose: Fix signedness issues wrt. digi count.
NET: SGIseeq: Add support for Challenge S Mezz board
do_exit(): make sure that we run with get_fs() == USER_DS
MTD: Add AMD Alchemy Mirage mapping driver.
econet: fix CVE-2010-3848
econet: disallow NULL remote addr for sendmsg(), fixes CVE-2010-3849
econet: fix CVE-2010-3850
NET: hdlcdrv: Make driver stop method stop queue.
CHAR: New key map for IBM WorkPad z50.
irda: Fix parameter extraction stack overflow
drivers/char/ppdev.c: fix information leak to userland
ipc: shm: fix information leak to userland
net: atm/clip does not work anymore as a module
ext3: ext3_symlink should use GFP_NOFS allocations inside
net: packet: fix information leak to userland
usb: core: fix information leak to userland
block: ioctl: fix information leak to userland
Change VERSION to 2.4.37.11

Summary of changes from v2.4.37.9 to v2.4.37.10
============================================
jfs: don't allow os2 xattr namespace overlap with others
sctp: Fix skb_over_panic resulting from multiple invalid parameter errors (CVE-2010-1173) (v4)
irda: Sock leak on error path in irda_create.
FAT: do not continue in fat_get_block if bmap fails
drivers/tun: MTU change for TUN/TAP interfaces
net: permanent NUD pins ethernet interfaces when ATM is compiled in.
sctp: fix to calc the INIT/INIT-ACK chunk length correctly is set
SCTP: Fix to encode PROTOCOL VIOLATION error cause correctly
irda: correctly free memory on irda_bind() failure
Change VERSION to 2.4.37.10

[themamboman]

Moving forward from 2.4 to 2.6 kernel may be very complicated as the driver kernel model has changed. I think Gary Colville found it easier to backport 2.6 drivers to 2.4. I'm going to see what it will take to do both.

I also found out that 3.x and 4.x aren't major changes from 2.6, just some numbering issues to avoid large numbers of version numbers after 2.6.

[themamboman]

Using your tarball exactly as it was (no config changes), the kernel and drivers built fine. However, using make modules_install had an error. Have you seen this?


depmod: *** Unresolved symbols in /lib/modules/2.4.37.11/kernel/drivers/net/e1000/e1000.0
depmod: dev_open

The file showed up though.

Is anyone interested in testing the output?

[SnkBitten]

I can't recall. I started building that module in the kernel, instead of as a module, as it's one of the devices that Virtualbox presents to the guest so I don't have to load a module file to have networking functioning in my VB Amithlon.

[shaf]

In the next few day I'll be testing an AM2 Motherboard (ASUS M3N-HT) using onboard Video. If you could add some notes on configuring the Network Component that would be helpful.

[SnkBitten]

Originally Posted by shaf

In the next few day I'll be testing an AM2 Motherboard (ASUS M3N-HT) using onboard Video. If you could add some notes on configuring the Network Component that would be helpful.

I need to rebuild my older setup so I'll take notes on what I do, probably doing that this weekend.

[SnkBitten]

Installation Guide Part 2 is in progress. Kernel 4 and drivers and installing Genesis and basic configuring are done. I'll be adding Picasso96Mode and a few other things at some point....

What's done is up with more to come.

http://amithlon.snkbitten.com/

[SnkBitten]

Oh yeah..I compiled Gary's kernel 3.00-test1 from the source he had made available. Wanting to get a feel for what he and Milan had done differently. Anyone have Gary's 3.10 source? That would be awesome to get my hands on.....

[shaf]

Just tried a test to boot with an AMD Athlon X2 250 CPU on an ASUS M3N-HT Motherboard with onboard Nforce 780a. Doesn't boot with Kernel 310 or Kernel 4. I'll have to drop down to an older motherboard.

[SnkBitten]

What was the setup? SATA or IDE - try disabling SATA or set to Legacy mode.
How much ram? - Start at 512 MB and add another for dual channel if desired..but start at 512 MB.
What type video? Did you try onboard video or a PCI video card? Bernd Meyer said on one forum that the video memory of the card gets mapped into the Amiga address space as well so it needs to be able to fit. Stick with 128 MB or 512 MB max for video ram.

The NVIDIA nForce 780a chipset could be the problem, but I'd try disabling SATA (or set to legacy), disabling onboard video and use a PCI supported card (they are cheap on eBay) and see if it works.

What did you get? Bouncing ball or black screen?

[shaf]

I have SATA disabled, using 512MB config and only onboard Video. Set 128Mb for Video ram in BIOS.
I get the bouncing ball and system freezes just after video config.
Next test I'll try with a PCI Video Card next and disable onboard VGA.

[SnkBitten]

It could be the vga= setting in the boot config (amithlon or small) is specifying a mode your onboard video doesn't support.

duh...never mind, you are seeing the bouncing ball so that vga= mode is working.

[themamboman]

Originally Posted by shaf

I have SATA disabled, using 512MB config and only onboard Video. Set 128Mb for Video ram in BIOS.
I get the bouncing ball and system freezes just after video config.
Next test I'll try with a PCI Video Card next and disable onboard VGA.

Set console_output to 4 and see if you can post the output (just a picture of where it freezes with your camera).

[croupier]

Quick question....anyone managed to get the Amithlon install CD transferred to a USB key (to make it bootable from a USB key; on devices without the CD/DVD unit)?

[themamboman]

Originally Posted by croupier

Quick question....anyone managed to get the Amithlon install CD transferred to a USB key (to make it bootable from a USB key; on devices without the CD/DVD unit)?

I used Rufus (windows app) with an iso from the Amithlon CD and made a bootable USB.

[croupier]

Thanks....guess will have to make another ISO image...Rufus is not accepting the one I have (raw mode 2)....