x1000 documentation and other x1000 related questions

@kas1e

"[HELO][DRAM]"
I had that error twice in 10 years. Fixed it by reconnecting my SATA cables to motherboard.

@sailor
Quote:

Can you post plz exactly information about your memory modules (full chip info, etc), and how much of them and in which ports, etc ,etc with which you have no such issue ?

As it start to looks like that issue is with the memory modules then...

@kas1e

Rock solid here, i use 2 sticks here.

Here has times when computer did boot, or reports 0 RAM, that’s when its time to replace backup battery. I have replaced PSU few times, the one I have looks like is quality and has lasted for a long time.

I have waster cooled my CPU, original CPU fan too noisy.
(I don’t think it has anything to do with it.)

@LiverForIt
Yeah, when i see this "[HELO][DRAM]" thing i firstly think about a battery, and about memory sticks not being fit correctly. But seems for one it fixed by repluging sata cables, for other one by swapping memory sticks.. strange

Yep I had this issue, I begin to think it just black magic sometimes as to weather you get a smooth working system or not, but i got the "Failed to load amigaboot.of: Insufficient memory" line a few times, I replace the sata cables to these https://www.amazon.co.uk/CORSAIR-Premi ... F&pd_rd_i=B07L687D8T&th=1

and it seem to fix the issue, it so weird.

I have 2 sticks and just 2Gb Ram that came with the system.

@kas1e
Quote:
I've never seen this issue on my x1000. I obtained mine new from Amigakit back when they were first released. I ordered mine with an RS232 port at the time so that was not an issue. My x1000 is mostly original except for a the RedeonHD 7750 and an RTL8169 PCI Gigabit Ethernet Controller I swapped in. It's still working fine. I just clean it a couple times a year. Oh, I did have to replace the battery a couple of times. Hope I didn't just jinx it.

I guess it would be nice if the on board Ethernet worked but it hasn't really been a poroblem. It would be nice if they could change CFE to better support the newer RadeonRX cards. I've thought about upgrading to an A5000 since my only backup right now is a uA1 but the cost is a bit to much for what I might gain at this point. I'm hoping to get a 1222 as a backup some day.

I have an Amiga X1000 on the way. Just wondering, is there a way to install the OS without having to use a CD-ROM drive? Maybe through the CF slot on-board or a USB stick?

Thanks!

@Nostromo
Sadly no USB boot :( There is a bug in the CFE or amigaboot: cfe can see usb, cfe can load amigaboot from usb, but then amigaboot didnt see usb-installcd as boot device. I was told that this is a cfe bug, but it looks like amigaboot didnt attach boot device tree properly.

As for cf card: i didnt tried yet, but as you can read on previous pages of topic TiredOfLife back in years use ide2cf adapter, and attach this adapter to ide bus of x1000 (yeah, x1000 still have one ide port and not only sata ones) and was able to boot installation cd from. But that need proper checking. I hope to check this myself soon

Thanks :)

@Nostromo

But if it helps, I believe you can use an external USB-connected CD-ROM drive.

Best regards,

Niels

@nbache

I'm pritty sure you can't

@LiveForIt

Quote:

It's good for keyboards where the assigned Amiga key is next to space and alt on the outside. Which is the proper Amiga layout. Some people swap the keys so the codes are sent backwards.

The only issue is they purposely broke it in the OS modules later on and removed support for it in Kickstart. So you can have a normal working keyboard in CFE but then if you need early startup menu or a boot shell you get problems. Like you press soft reset only to find it does a hard reset! Because the keys aren't swapped. For some reason they took support of this out of the OS without thinking it seems. I don't think anyone was using this configuration when they took it out or they surely would have noticed that CFE, AmigaDOS and Workbench are using different keys for Alt.

@kas1e

Quote:

I found when I first used it CFE would easily crash when exiting AmigaBoot. They weren't going to fix it or didn't know how so CFE would remain in an unstable beta state. Funny now the AEON version doesn't let you exit.

Oh here's another one. Cursor up in CFE. History is good to have but all the past command merging looks confusing.

Quote:

The X1000 hardware can be sensitive. I found even a thin power cable can cause it to crash and reboot. I had to swap my power cable for a thicker one in my collection when I took my X1000 out for a walk at the Amiga club and and had problems.

I've seen the insufficient memory confusing error many times. And when it did boot sometimes it would stall. Or would go into a crash and reboot loop. Needing hard reset on case to be pressed or it wouldn't follow through.

It started to get more unstable as time went on. It started crashing when booting from HDD. Then needed power cycle or another reboot before it booted fully. At one staged I swapped around my RAM sticks which seemed to help, or it just cleaned the slots, since PPC doesn't like dirt and grime build up. Then I had power problems. Didn't want to turn off. Black screen of death etc.

So I replaced the battery, replaced the RAM and put in two 2GB sticks interleaved for optimal access, and replaced my SATA cables with proper ones with holding clips. It's been good ever since. Does my head in with sloppy firmware that just crashes instead of verifying and giving an error.

That reminds me, the X1000 board should have a graphics card holding clip, but they didn't put one on so in some positions your graphics card can slip out of place.

Quote:

I think they need to come up with a better excuse. As you found at least one AmigaBoot has support for USB booting in the change log. Still need to track that sucker down. CFE can boot from USB. I booted Linux kernels for years off a USB stick plugged into my Mac keyboard. I think the problem is AmigaBoot.

Likely I didn't run into too many USB issues because of this. Mac keyboards have a built in hub and it is plugged into keyboard port. So CFE could easily see USB sticks almost accidentally because of my setup. I tried in other ports and CFE did have some detection issues. Most of the time it detected but wouldn't mount.

However my case does have a USB bug. I have ThermalTake VF-I which has USB2/3 ports on front I plugged into board headers. The front USB2 port has never worked and crashes OS4. The USB3 port works fine. Never solved that.

@daveyw

Quote:

I don't think that's the reason. They still have their documentation and beta driver. It works fine on Linux and they can compare code bases to see any obvious issues. Why has it worked on Linux fine for years and never worked fully on OS4 I wonder?

In any case I think the OS4 RTL drivers have issues. Compared to my XE my X1000 network is slow. My XE can max out my internet but my X1000 doesn't. Linux runs at full speed on both network adapters. Internal and RTL8169 PCI I think it is.

@ddni

Quote:

I couldn't do that hack. I read all sorts of issues initially and I'm not interested in more issues than I already have with my R7 250. Aside from that I couldn't use CFE blind as I need to see my boot menu. Needing to count cursor moves would be like the time I needed to put my phone into recovery mode by pressing the right keys in sequence. When my screen was broken and a false move would wipe it.

Quote:

In my experience, the device side is always well documented. You might have to get used to the style of writing but with an example linux driver at hand, it is often self-explaining.

In my limited experience with the amiga platform, it's the amiga side that is the issue. There is a notorious lack of information of how things are done.

The only option is reinventing the wheel based on trial and error. Which often leads to a driver with only the basic functionallity implemented. (it works so it's done)

For example:
I've just started to investigate the possibility of a new driver for private usage. The device protocol standard is easy. I can have this up and running over the weekend. The amiga side, which should be trivial as well (it's one of the core components of a computer), is lacking information on all levels.

It starts with a simple explaination of how to write an aos4 device in the first place. All explaining and example code assumes a certain level of amiga coding knowledge which amiga coding noobs like me don't have.

Then there's the lack of information of which commands needs to be implemented for a certain type of device.

So I have to create a dummy driver first to probe which commands are called by applications and other components before I can get started. (Inventing the wheel)

This is very frustrating for the generation X who just want to get things done.

@geennaam
Can it be you who can wrote Ethernet onboard driver for x1000 ?:) Your driver for audio cards is 100% stable and good (i use it on 3 different amigaos4 machines for now). If you will consider ethernet driver i can help with betatesting of course

Anyway, ppls who wrote drivers offten didn't wrote any documentation, so other ones need to do the same again and again. And once they do what they want, they then have no big interest or motivation to wrote documentation, and so other devs recreating the wheel again.

@geennaam

Quote:
Speaking of which - would you consider the possibility to transfer/license your audio driver to OpenAmiga, so that other developers could learn from your code how to write an AHI driver for OS4?

@Hypex
Quote:

Interesting. Yes, back when I had an RTL8169 in my X5000 I moaned how slow it was under OS4 vs the same hardware under Linux. But the RTL8169 was a temperamental beastie and eventually stopped working so I switched to the onboard ethernet.

Sadly my XE no longer boots so I can't check the speed there.

@kas1e
Quote:
Then I will not only face all of the problems described above. But in addition also face the issues that I do not own a X1000 and don't have access the required documentation for the PAsemi processor (probably NDA material which is only available the Hyperion, a-eon and kernel developers). So this will be a no-go.

Quote:
The only reason that I can think of why my driver performs as it does is because I use an AHI sample timing method that totally makes sense from audio and PCIe (chip) perspective but its existance is only mentioned in two sentences in the AHI developer documentation. Also none of the open source AHI drivers that I could find use this method. They all use the method as if DMA busmasters don't exist. So basically I had to find out how to implement this AHI timing method with trial and error as well

Quote:
In my opinion, this is not the responsibility of third party driver developers but instead should be part of the SDK documentation.