New EHCI driver fails on my system

A forum for AmigaOne XE and MicroA1-C specific issues.
Evillord68
Beta Tester
Beta Tester
Posts: 233
Joined: Sun Jun 19, 2011 5:26 pm

Re: New EHCI driver fails on my system

Post by Evillord68 »

paulsy wrote:after hours of experimenting i have found that the EHCI driver does not freeze the system if I remove Devs/Networks/eth3com.device.
put it back in and the problem comes back, so it seems to conflict with the ethernet for some reason.
at least I know whats causing it even if i dont know why :)
I will post this in Bugzilla.

Greetings

Jürgen
User avatar
Hypex
Beta Tester
Beta Tester
Posts: 645
Joined: Mon Dec 20, 2010 2:23 pm
Location: Vic. Australia.

Re: New EHCI driver fails on my system

Post by Hypex »

Can you give more PCI info paulsy? Perhaps with PCIScan on OS4Depot. This looks to be the same device. Do you have the same Vendor code? And other things like Base address?

I have tested the following card yesterday in a friends A1 as working. DiskMonTools reported an average of 6.5MB/s on the flash drive I tested at the time.

The box calls it a Sunix UP205. It has an NEC chipset. 5 port (4+1) USB. The PCI card itself is of a yellow kind of colour and doesn't look very nice. But it works better than it looks! :-)
Hidden Text - Click to Show :
Device #11:
Vendor: 0x1033 (NEC Corporation)
Device: 0x0035 (USB)
Class code: 0x0C0310
RevisionID: 0x43
Interrupt: Line 0x0A, Pin 0x01
Subsystem Vendor: 0x1033
Subsystem ID: 0x0035
Resource range #0 (MEMORY)
Base address: 0x88101000
Physical address: 0x88101000
Size: 0x00001000

Device #12:
Vendor: 0x1033 (NEC Corporation)
Device: 0x0035 (USB)
Class code: 0x0C0310
RevisionID: 0x43
Interrupt: Line 0x0B, Pin 0x02
Subsystem Vendor: 0x1033
Subsystem ID: 0x0035
Resource range #0 (MEMORY)
Base address: 0x88102000
Physical address: 0x88102000
Size: 0x00001000

Device #13:
Vendor: 0x1033 (NEC Corporation)
Device: 0x00E0 (USB 2.0)
Class code: 0x0C0320
RevisionID: 0x04
Interrupt: Line 0x07, Pin 0x03
Subsystem Vendor: 0x1838
Subsystem ID: 0x1074
Resource range #0 (MEMORY)
Base address: 0x88103000
Physical address: 0x88103000
Size: 0x00000100
paulsy
Posts: 6
Joined: Sun Sep 04, 2011 5:11 pm

Re: New EHCI driver fails on my system

Post by paulsy »

I used pciscan to get the information you requested.there are some differences and its not yellow.
Hope you find this helpfull.

Thanks

Device #12:
Vendor: 0x1033 (NEC Corporation)
Device: 0x0035 (USB)
Class code: 0x0C0310
RevisionID: 0x43
Interrupt: Line 0x09, Pin 0x01
Subsystem Vendor: 0x167E
Subsystem ID: 0x0035
Resource range #0 (MEMORY)
Base address: 0x88101000
Physical address: 0x88101000
Size: 0x00001000

Device #13:
Vendor: 0x1033 (NEC Corporation)
Device: 0x0035 (USB)
Class code: 0x0C0310
RevisionID: 0x43
Interrupt: Line 0x0A, Pin 0x02
Subsystem Vendor: 0x167E
Subsystem ID: 0x0035
Resource range #0 (MEMORY)
Base address: 0x88102000
Physical address: 0x88102000
Size: 0x00001000


Device #14:
Vendor: 0x1033 (NEC Corporation)
Device: 0x00E0 (USB 2.0)
Class code: 0x0C0320
RevisionID: 0x04
Interrupt: Line 0x0B, Pin 0x03
Subsystem Vendor: 0x167E
Subsystem ID: 0x227A
Resource range #0 (MEMORY)
Base address: 0x88103000
Physical address: 0x88103000
Size: 0x0000010
Evillord68
Beta Tester
Beta Tester
Posts: 233
Joined: Sun Jun 19, 2011 5:26 pm

Re: New EHCI driver fails on my system

Post by Evillord68 »

Evillord68 wrote:
paulsy wrote:after hours of experimenting i have found that the EHCI driver does not freeze the system if I remove Devs/Networks/eth3com.device.
put it back in and the problem comes back, so it seems to conflict with the ethernet for some reason.
at least I know whats causing it even if i dont know why :)
I will post this in Bugzilla.

Greetings

Jürgen

Today i have send a debugoutput to Bugzilla. Sorry for the long waiting.
I do not have the cable for output, before.
Now i use a normal PC cable on onboard serial port, it works.

Greetings

Jürgen
User avatar
Raziel
Posts: 1170
Joined: Sat Jun 18, 2011 4:00 pm
Location: a dying planet

Re: New EHCI driver fails on my system

Post by Raziel »

Evillord68 wrote:
Evillord68 wrote:
paulsy wrote:after hours of experimenting i have found that the EHCI driver does not freeze the system if I remove Devs/Networks/eth3com.device.
put it back in and the problem comes back, so it seems to conflict with the ethernet for some reason.
at least I know whats causing it even if i dont know why :)
I will post this in Bugzilla.

Greetings

Jürgen

Today i have send a debugoutput to Bugzilla. Sorry for the long waiting.
I do not have the cable for output, before.
Now i use a normal PC cable on onboard serial port, it works.

Greetings

Jürgen
You probably won't get any output through serial anyway - i tried...

I just tested my VIA based (the only one i got) USB2 PCI card and while it is plugged in with no hubs (or with a HAMA 2.0 USB hub, doesn't really matter) or Mass Storage devices on it works, nothing happens...even with ehci loaded and shown as USB2 in USBInspector.

As soon as i put on a (powered USB2) hub with a Mass Storage device, it won't take long and it will freeze the whole system, just as @paulsy described (no serial, no crash, no oputput whatsoever).
Even the reason for it to freeze is completely random, sometimes while browsing, sometimes while simply clicking on a window to activate it, even while it read/writes to/from the USB Mass Storage that is plugged in.

Here is the pciscan output of the USB card in question.
One thing that is strange is that it says "MEMORY" at "Resource Range" while with all the other USB1.1 devices it says "IO", but that is probably normal?
Hidden Text - Click to Show :
Device #14:
Vendor: 0x1106 (VIA Technologies, Inc.)
Device: 0x3104 (USB 2.0)
Class code: 0x0C0320
RevisionID: 0x63
Interrupt: Line 0x09, Pin 0x03
Subsystem Vendor: 0x1106
Subsystem ID: 0x3104
Resource range #0 (MEMORY)
Base address: 0x98102000
Physical address: 0x98102000
Size: 0x00000100
These are the other USB1.1 devices i have in my system and all of them work perfectly
Hidden Text - Click to Show :
Device #6:
Vendor: 0x1106 (VIA Technologies, Inc.)
Device: 0x3038 (VT82xxxxx UHCI USB 1.1 Controller)
Class code: 0x0C0300
RevisionID: 0x1A
Interrupt: Line 0x05, Pin 0x04
Subsystem Vendor: 0x0925
Subsystem ID: 0x1234
Resource range #4 (IO)
Base address: 0x00802080
Physical address: 0x00802080
Size: 0x00000020

Device #7:
Vendor: 0x1106 (VIA Technologies, Inc.)
Device: 0x3038 (VT82xxxxx UHCI USB 1.1 Controller)
Class code: 0x0C0300
RevisionID: 0x1A
Interrupt: Line 0x05, Pin 0x04
Subsystem Vendor: 0x0925
Subsystem ID: 0x1234
Resource range #4 (IO)
Base address: 0x008020A0
Physical address: 0x008020A0
Size: 0x00000020

Device #12:
Vendor: 0x1106 (VIA Technologies, Inc.)
Device: 0x3038 (VT82xxxxx UHCI USB 1.1 Controller)
Class code: 0x0C0300
RevisionID: 0x61
Interrupt: Line 0x0B, Pin 0x01
Subsystem Vendor: 0x1106
Subsystem ID: 0x3038
Resource range #4 (IO)
Base address: 0x00802500
Physical address: 0x00802500
Size: 0x00000020

Device #13:
Vendor: 0x1106 (VIA Technologies, Inc.)
Device: 0x3038 (VT82xxxxx UHCI USB 1.1 Controller)
Class code: 0x0C0300
RevisionID: 0x61
Interrupt: Line 0x07, Pin 0x02
Subsystem Vendor: 0x1106
Subsystem ID: 0x3038
Resource range #4 (IO)
Base address: 0x00802520
Physical address: 0x00802520
Size: 0x00000020
I need to add that this is on an upd4 installation.

Thank you

EDIT:
Testing without the hub - no change
Testing with no eth3com.device - no change

I think it may be EHCI only that does this, because if no IO is done over an installed EHCI hardware with ehci loaded in kickstart, it will work just fine :-)
People are dying.
Entire ecosystems are collapsing.
We are in the beginning of a mass extinction.
And all you can talk about is money and fairytales of eternal economic growth.
How dare you!
– Greta Thunberg
User avatar
Raziel
Posts: 1170
Joined: Sat Jun 18, 2011 4:00 pm
Location: a dying planet

Re: New EHCI driver fails on my system

Post by Raziel »

I got myself a Vivanco 23418 PCI IO 4+1 USB card with NEC chip
Hidden Text - Click to Show :
Device #12:
Vendor: 0x1033 (NEC Corporation)
Device: 0x0035 (USB)
Class code: 0x0C0310
RevisionID: 0x43
Interrupt: Line 0x0B, Pin 0x01
Subsystem Vendor: 0x1033
Subsystem ID: 0x0035
Resource range #0 (MEMORY)
Base address: 0x98102000
Physical address: 0x98102000
Size: 0x00001000

Device #13:
Vendor: 0x1033 (NEC Corporation)
Device: 0x0035 (USB)
Class code: 0x0C0310
RevisionID: 0x43
Interrupt: Line 0x07, Pin 0x02
Subsystem Vendor: 0x1033
Subsystem ID: 0x0035
Resource range #0 (MEMORY)
Base address: 0x98103000
Physical address: 0x98103000
Size: 0x00001000

Device #14:
Vendor: 0x1033 (NEC Corporation)
Device: 0x00E0 (USB 2.0)
Class code: 0x0C0320
RevisionID: 0x04
Interrupt: Line 0x09, Pin 0x03
Subsystem Vendor: 0x1033
Subsystem ID: 0x00E0
Resource range #0 (MEMORY)
Base address: 0x98104000
Physical address: 0x98104000
Size: 0x00000100
I can refrain from my previous post, nothing changed.
Although it seems this card is doing more in the background (like retaining the copy operation), because the lights on the USB sticks still flicker and i get a Read/Write error warning window when i plug one out (but this is already when WB is completely stuck and unresponsive and not reacting to anything i try to do ... other operations, like scrolling text in a shell is frozen and probably crashed.

Workbench has to this point killed off all IO (mouse, keyboard) no matter if USB or PS/2 which is really bad.
One of my clocks (the one in AmiDock) keeps running, the other (FlipClock on WB) is stalled and won't come back.

It only takes one big copy (unarc) operation with two USB sticks plugged in the card (unarc from one USB stick to the other) to make WB (IO) stall completely.

The hardware seems to work (both VIA based and NEC based) but the EHCI driver is probably doing bad things :-(

EDIT: Further testing shows that if the card is in place and EHCI is loaded it will bring down my system after some time even if nothing is put into the USB ports...that is even more bad :-(
People are dying.
Entire ecosystems are collapsing.
We are in the beginning of a mass extinction.
And all you can talk about is money and fairytales of eternal economic growth.
How dare you!
– Greta Thunberg
User avatar
Hypex
Beta Tester
Beta Tester
Posts: 645
Joined: Mon Dec 20, 2010 2:23 pm
Location: Vic. Australia.

Re: New EHCI driver fails on my system

Post by Hypex »

Raziel wrote:I can refrain from my previous post, nothing changed.
This is strange but I have read of NEC cards failing before also. And your symptoms look like mine. But I get read errors when the USB key has been plugged in a while and endlessly flashing it's light. And also when pulling out. I had some success with a EMTEC/BASF 16GB key and couldamost write files to it but it always broke eventualy.

I even thought it looked ike the USB MaxTransfer was set too high looking at the evidence so looked for a way to lower it. LOL. :-)

BTW, have you tested it on a fresh OS4.1.5 partition? I setup a fresh install of the latest update which I leave untouched for testing just ot make sure it isn't my setup.
User avatar
Raziel
Posts: 1170
Joined: Sat Jun 18, 2011 4:00 pm
Location: a dying planet

Re: New EHCI driver fails on my system

Post by Raziel »

Hypex wrote: BTW, have you tested it on a fresh OS4.1.5 partition? I setup a fresh install of the latest update which I leave untouched for testing just ot make sure it isn't my setup.
Nope.

I have put the USB2 testing on my XE on ice.
I need internet but can live without USB2 for as long as the X1000 needs to arrive ;-)
People are dying.
Entire ecosystems are collapsing.
We are in the beginning of a mass extinction.
And all you can talk about is money and fairytales of eternal economic growth.
How dare you!
– Greta Thunberg
amigaonefan
Posts: 43
Joined: Fri Sep 16, 2011 9:20 pm

Re: New EHCI driver fails on my system

Post by amigaonefan »

@Raziel

I have a card for my microA1 that uses the exact same VIA USB chip 0x3104 device and rev 0x63 and I see the exact same problems you do with the EHCI module. If I comment out EHCI from kickstart/kicklayout and reboot, my 4 port USB 2.0 card works, but of course, only in USB 1.1 mode.
amigaonefan
Posts: 43
Joined: Fri Sep 16, 2011 9:20 pm

Re: New EHCI driver fails on my system

Post by amigaonefan »

Some more information....

The multi card 4 port USB 2 using VIA chipset under OS4.0 final works correctly with no lockups even when the internal VIA usb ports are both enabled. This differs from OS4.1 update 5, in which I have to disable the internal usb ports or else the system locks up as soon as I insert a usb device in the multi card. It seems then that something in the usb 1.1 code has changed between Os4.0 final and Os4.1 update 5 (in addition to adding the USB 2 layer) that is causing conflict with the internal USB VIA chips on the micro A1.
Post Reply