"Happy New Year" 3124 Resolution build, Freeware ver. 5.3.2b2 ("HK Pirate")

[(S)ATAman]

The "Happy New Year Resolution" build for 3124 is attached (again).

Surprise: it's a very early build and very slow.
I tried with an SSD - it's much slower than the corresponding build made by Vu from Silicon Image.

The only reason anyone would use it is the bootability on G4 / G5.

It is also based on API-s which later became obsolete (Kernel-User Notification), if you hot-swap the drives you notice the dialogs.
Later builds are better, but as a starting point I used the same version what is pirated in Hong Kong.

Instructions:

You will see three files:

- SeriTek2SE4Flasher, this works on PowerPC only.
- FT_ATA_Sil312E.kext
- Sil3112DeviceNub.kext

First, we need to tell the system that there is an ATA device nub.
The simplest:

1) sudo cp -r  /Volumes/SeriTek2SE4Flash/Sil3112DeviceNub.kext /tmp
(the system will ask for admin password)
2) sudo cp -r /tmp/Sil3112DeviceNub.kext /System/Library/Extensions

Loading the kernel extension temporarily:
3) sudo cp -r /Volumes/SeriTek2SE4Flash/FT_ATA_Sil3132E.kext /tmp
4) sudo kextload /tmp/FT_ATA_Sil3132E.kext

After few seconds the Silicon Image 3124 controller is operational.
Do not attach any drives to it, proceed with flashing.

As it finishes, you are ready and are bootable on the "X".

[macStuff]

bootable on OSX but not os9?
please, we need to be very clear about such things

[(S)ATAman]

bootable on OSX but not os9?
please, we need to be very clear about such things

Not yet as it wasn't done ever.

The first thing is to update the "X" drivers and remove the ROM lock.
The "9" comes after that. No one ever did it, so it has to be written from scratch.

[(S)ATAman]

To be clear: neither SImage nor FirmTek did develop a single bit of the code for 3132 / 3124 (they are similar!) in last probably 12 years.
I can testify that what I did 12 years ago was some fix for the ExpressCard only.
The base code was unmodified probably since 2006 - 2007. And that is only the FCode (from me) and the driver ("X" only, either from me or from SImage, no other choice exists).

In these 13 years much changed, the existing code is barely sufficient today. Both my code and the code from SImage panics.
I panic on hot-swap, SImage (their code from Monday, Sept. 10 2012) panics on simple I/O test.

The best performance both code can achieve in MDD G4 with a modern SSD is about 130 MB/Sec on write (me), 100 MB/Sec on write (SImage), 200 MB/Sec on read (me), 220 MB/Sec on read (SImage).

This is what we can expect on "9".

The root cause for the slowness is the design of the card (3124). Fixing, updating and the unlocking the existing "X" driver is a must before doing anything else. Luckily it is not that difficult, the tools today are far better than what they used to be.

And from today's perspective these 13 year old pieces of software we wrote are quite outdated. Looking at them gives few thoughts how a "9" part of it has to be done.