@ntromans

Thanks. For FAT32 issue, what was your setup?

Did you boot from CDROM and FAT32 pendrive was the only pendrive or did you boot from SFS pendrive and FAT32 pendrive was the second pendrive in the system?

[quote name=deadwood post=11489]@deadwood - @ntromans

Thanks. For FAT32 issue, what was your setup?

Did you boot from CDROM and FAT32 pendrive was the only pendrive or did you boot from SFS pendrive and FAT32 pendrive was the second pendrive in the system?[/quote]

The same result either when booting from CD (so FAT was the first pendrive on te system) or booting from AROS unstalled to USB.

Cheers,
Nigel.

for me first iso usb key works on my laptop acer aspire 5630

second iso usb key works

all best

The two ISO images were burned and successfully installed via VMware, then tested on two Dual-Core PCs

Test ISO exp-usb3-pc-x86_64-20260327
- The USB stick was burned via VMware with USB 2.0 configured
- AROS boots correctly when ‘noacpi’ is added.
- Second USB stick formatted as FAT32 is displayed in Wanderer but an error message appears first, see screenshot
- Some USB sticks formatted as FAT32 are not displayed in Wanderer

Test ISO pc-x86_64-20260325-014650
- Burned the USB stick via VMware configured for USB 1.1 (USB 2.0 generates an error during installation)
- AROS boots correctly when noacpi is added.
- The second USB stick formatted as FAT32 is displayed correctly, but some FAT32 USB sticks are not displayed on Wanderer

Thanks, that confirms that is some sort of regression with USB sticks handling. I added this to my TODO.

Hi,

On a USB2/USB3 hybrid laptop, with the usb3 iso if a usb device is in a usb2 socket and removed and placed into a usb3 socket the laptop trackpad stops working until the laptop is rebooted.
Not sure if it is the laptops' AMD usb "feature" (AMD does have quirks regarding their implementation of usb3) but does not happen with usb2 based iso, the trackpad continues to work.

@Jeff1138

Please run PCITool and select "Save Complete PCI information to RAM" and post the file here.

Thanks, copied PCITool to System: and was able to save file to RAM: but unable to close PCITool afterwards and save file will not copy to anywhere else as it is locked to PCITool which cannot be closed. Copied Scout over and used that to close PCITool but file is still locked.

Any ideas?

Just as an aside, I do get random freezes with PCITool, using the arrow keys to go up and down the list and around the mid point, it will freeze.

I've seen these PCITool freezes on some configurations but haven't been able to locate the source yet.

With regards to export: I exporting the file does not work for you try this:

In PCITool locate the USB xHCI device. Click "Save device information to ram" (not the "complete" button). This will save one devices information. Also note the value in the IRQ field. Now try to scroll through other devices and note if there is a device with the same IRQ information. Save it's information too.

Hi,

Can only find other HCI's as per attached. If it helps the random freezes only happen on 00.14.2 00.14.3 only in PCITool.

Ok, interrupts seem good from your export. Try starting with 'noioapic' kernel argument and see if the usb2->usb3 sequence still causes your trackpad to stop working.

Hi,

Thanks, just tried but still the same

New build:

[url]https://build.axrt.org/download/builds/Experimental/exp-usb3-pc-x86_64-20260331-170620.tar.gz[/url]

Fixes:
- Pendrives are now recognized and mounted under VirtualBox
- General fix for interrupt routing under VirtualBox

Known issues:
- Second USB stick does not work properly

As a reminder, if you want to support Kalamatee's work you can:
- Send him a direct donation to: kalamatee@gmail.com
- Donate through Ko-Fi: [url]https://ko-fi.com/kalamatee[/url]


Here is a summary of currently tested USB3 support:

VirtualBox 7.2:
- mount FAT32 pendrive - OK
- copy 1G data from FAT32 to RAM - OK
- InstallAROS to SFS pendrive - OK

QEMU:
- mount FAT32 pendrive - OK
- copy 1G data from FAT32 to RAM - FAIL (crashes system)
- InstallAROS to SFS pendrive - OK

VMWARE 17:
- mount FAT32 pendrive - OK
- copy 1G data from FAT32 to RAM - OK (only if VM RAM < 4GB )
- InstallAROS to SFS pendrive - FAIL

Following hardware has been tested to boot with live pendrive in USB3 port:
- VirtualBox
- QEMU

Following hardware does not boot with live pendrive in USB3 port:
- VMWARE
- Dell Latitude E7450
- Dell Latitude 7390
- Fujitsu Futro S720
- GA-H81M-S2V
- ASUS P8Z68V LX

With this new ISO (usb3-pc-x86_64-20260331-170620) on VMware 12 using USB 2.0 and USB 3.0, the installation was successful.

When the USB stick was booted on a native PC USB 2.0, the second USB stick, formatted as FAT32, displayed file and folder names containing strange characters.

With the latest build above, I can't boot at all on VirtualBox 6.1. Screenshot 1 is the default boot option; Screenshot 2 is with noacpi added.

The VM had a USB 2 controller, but turning off USB completely didn't change anything.

Boot everything fine, FAT key on USB 2 is not recognized correctly, in the previous two ISOs it works regularly, my computer is an HP Z400

@Amiwell79

Thanks, noted issue on your hardware. I see similar issue on some of my hardware too.

@ncafferkey

Which of these ISOs is the last one that works on your VBox 6.1:

[url]https://build.axrt.org/download/builds/AROS-ABIv11/[/url]

Also there was an old report stating that IO-APIC had to be disabled in VirtualBox 5 configuration to boot AROS. This is no longer necessary with VirtualBox 7, but I don't recall about VirtualBox 6.

I bought a USB 3 card I will do some tests when it arrives

[quote name=deadwood post=11515]
@ncafferkey

Which of these ISOs is the last one that works on your VBox 6.1:

[url]https://build.axrt.org/download/builds/AROS-ABIv11/[/url]

Also there was an old report stating that IO-APIC had to be disabled in VirtualBox 5 configuration to boot AROS. This is no longer necessary with VirtualBox 7, but I don't recall about VirtualBox 6.[/quote]

I was actually able to boot the original USB3 ISO I had been trying once I enabled nested paging in the VM. Is this expected, or should I bisect to establish when that became necessary?

BTW, I think IO-APIC will be disabled whenever ACPI is disabled, and that was one of my original tests