@scabit

Quote:

scabit wrote: @Hypex Do you remember if the kernelhd-2.6.18 had the cache coherency issue A1 patch for USB applied to it?

The 2.6.18 kernel for the A1 has the non cache coherent DMA implementation enabled for any driver that uses the DMA API.

Quote:

This is what I think is going on...please correct me if any of this is wrong... It seems (from reading the USB 2 official specifications) that since EHCI uses DMA for data transfer, any machine which uses an EHCI driver and has cache coherency problems with dma access must have a patch or workaround in order for usb to work properly.

Right, the DMA API ensures cache coherency through the allocation of non-cacheable memory and/or explicit cache flushes.

Quote:

The micro A1 ( and I think teron A1SE and A1XE) boards all have this cache coherency issue related to the PowerPC 750 processor.

It's not related to the PPC750, but rather to the northbridge. From what I heard the Articia S northbridge does not always deliver snoop events to the CPU.

Quote:

This showed up first with the fact that the original uboot on the A1's did not allow anything but PIO for the onboard IDE, and any attempt at DMA for the onboard IDE resulted in many errors. This got fixed on the linux side with a patch (which can be found here on sourceforge. )

Yes, this is fixed by the normal AmigaOne patch, which defines NOT_COHERENT_CACHE to compile in the above mentioned DMA implementation.

Quote:

Similarly, there is a patch for the EHCI USB and audio (ASLA) components in linux as well.

The EHCI patch fixes a different issue: on the PowerPC architecture atomic variables cannot be placed in non cacheable memory, because the PPC architecture uses a load/store reservation scheme for atomic operations and reservations are only allowed for variables in cacheable memory. Unfortunately the EHCI driver in the 2.6.18 kernel violated this rule.
(PPC experts may correct me, if I'm wrong.)

The ALSA patch works around an API issue, because ALSA maps DMA memory to userspace. The patch is actually a very bad hack, because it forces the ALSA sound drivers to allocate normal cacheable memory for the DMA transfers to/from the sound card (and thus bypasses the DMA API). This "works" most of the time, because the northbridge only looses snoop requests from time to time.

Quote:

I am looking for a 2.6.18 kernel that has these patches installed, plus the VIA SATA Driver incorporated so I can test a combo USB/SATA PCI card in my micro AmigaOne. In OS4.1 update 5, the USB ports in this new card work...but only if I disable the EHCI module in the OS4 kernel....which makes me suspect this same EHCI cache coherency issue is the culprit in OS4 as well as linux in getting USB2 to work.

All the A1 Linux v2.6.18 kernels should have these patches applied. Can you show me the output of "lspci -vvv" and dmesg? Otherwise it is hard to tell, why your USB2 card is not working.

Quote:

Now, the onboard USB on a micro is only USB 1.1, I know, but from what I can understand, even using USB1.1 with a driver written for USB 2 (and thus having an EHCI component) causes problems on a system with cache coherency DMA issues unless those issues are addressed.

USB 1.x and USB 2.0 functionality is handled by different drivers, even if the functionality is provided by the same controller chip. ohci_hcd, uhci_hcd are for USB 1.x and ehci_hcd for USB 2.0.

Quote:

The kernell linux patch for USB should fix my USB issues in linux. If it does not, there still needs to be some work done to get this cache coherency problem fixed. But I am not sure yet if I have used a linux kernel that has the USB patch applied.

Sure, this non cache coherent DMA implementation has issues on PPC CPUs (multiple virtual memory mappings for the same physical address with different page attributes), but usually it works good enough.

Edit: patch descriptions

Last edited by Geri on 01-Dec-2012 at 09:30 PM.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri

Here is the dmesg stuff relating to my usb card.....

libata version 2.00 loaded.
ehci_hcd 0000:00:08.2: EHCI Host Controller
ehci_hcd 0000:00:08.2: new USB bus registered, assigned bus number 1
ehci_hcd 0000:00:08.2: irq 11, io mem 0x88100100
ehci_hcd 0000:00:08.2: USB 2.0 started, EHCI 1.00, driver 10 Dec 2004
usb usb1: configuration #1 chosen from 1 choice
hub 1-0:1.0: USB hub found
hub 1-0:1.0: 4 ports detected
ohci_hcd: 2005 April 22 USB 1.1 'Open' Host Controller (OHCI) Driver (PCI)
USB Universal Host Controller Interface driver v3.0
uhci_hcd 0000:00:07.2: UHCI Host Controller
uhci_hcd 0000:00:07.2: new USB bus registered, assigned bus number 2
uhci_hcd 0000:00:07.2: irq 5, io base 0x00802080
usb usb2: configuration #1 chosen from 1 choice
hub 2-0:1.0: USB hub found
hub 2-0:1.0: 2 ports detected
uhci_hcd 0000:00:07.3: UHCI Host Controller
uhci_hcd 0000:00:07.3: new USB bus registered, assigned bus number 3
uhci_hcd 0000:00:07.3: irq 5, io base 0x008020a0
usb usb3: configuration #1 chosen from 1 choice
hub 3-0:1.0: USB hub found
hub 3-0:1.0: 2 ports detected
uhci_hcd 0000:00:08.0: UHCI Host Controller
uhci_hcd 0000:00:08.0: new USB bus registered, assigned bus number 4
uhci_hcd 0000:00:08.0: irq 9, io base 0x00802400
usb usb4: configuration #1 chosen from 1 choice
hub 4-0:1.0: USB hub found
hub 4-0:1.0: 2 ports detected
uhci_hcd 0000:00:08.1: UHCI Host Controller
uhci_hcd 0000:00:08.1: new USB bus registered, assigned bus number 5
uhci_hcd 0000:00:08.1: irq 10, io base 0x00802420
usb usb5: configuration #1 chosen from 1 choice
hub 5-0:1.0: USB hub found
hub 5-0:1.0: 2 ports detected
Initializing USB Mass Storage driver...
usbcore: registered new driver usb-storage
USB Mass Storage support registered.
usbcore: registered new driver hiddev
usbcore: registered new driver usbhid
drivers/usb/input/hid-core.c: v2.6:USB HID core driver
serio: i8042 AUX port at 0x60,0x64 irq 12
serio: i8042 KBD port at 0x60,0x64 irq 1
mice: PS/2 mouse device common for all mice
i2c /dev entries driver
TCP bic registered

Here is the lspci -vvv for the SYBA card USB and IDE/SATA controller

00:08.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61) (prog-if 00 [UHCI])
Subsystem: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-

Last edited by scabit on 29-Nov-2012 at 08:47 PM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Quote:

scabit wrote: @Geri Okay, I have your kernel with SATA support. I connect a SATA drive to my SYBA PCI card SATA port and boot the AmigaOne into Lenny. Here is what dmesg shows me about the SATA..... sata_via 0000:00:08.3: version 2.0 sata_via 0000:00:08.3: routed to hard irq line 9

Okay, interrupts are fine.

Quote:

ata1: SATA max UDMA/133 cmd 0x802440 ctl 0x80244A bmdma 0x802480 irq 9 ata2: SATA max UDMA/133 cmd 0x802450 ctl 0x80245A bmdma 0x802488 irq 9 scsi0 : sata_via ata1: SATA link down (SStatus 9F9F9F9F SControl 9F9F9F9F) ATA: abnormal status 0x8 on port 0x802447 scsi1 : sata_via ata2: SATA link down (SStatus 9F9F9F9F SControl 9F9F9F9F) ATA: abnormal status 0x8 on port 0x802457 PCI: Enabling device 0000:00:05.0 (0084 -> 0085) cannot find the slot for index 0 (range 0-1)

These 9F9F9F9F values look more like the driver is reading random data, but not the actual register in the PCI IO space. I wonder, if the hardware only support 16 bit PCI I/O addresses, as it is the case with the AC'97 controller in the VIA southbridge.

Quote:

I set the drive up previously by connecting it to a different IDE port (my SATA drive is really an IDE drive with an IDE to SATA converter on it), and partitioning it with two ext3 partitions. On each partition I copied a directory - videos on one, documents on another, so I could identify them if I neede to by content.

I also have a IDE drive attached to a sii3114 SATA controller through a SATA to IDE converter.

Quote:

So what does the listing above mean? Is there anything I can attempt to mount? I have tried mounting it like this: mount -t ext3 /dev/sda /mnt/hdb2 using everything from sda1 to sda8, sdb, etc...no success.

Well, the kernel couldn't probe any SATA device, so mounts will fail. But again, the output of "lspci -vvv" could be helpful.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri

here is the lspci -vvv...it was cut off due to size limitations in the previous message. HMm...no, there must be some embedded xontrol character that Amigaworld sees causing it to end the display. ...okay, had to remove all the greater than and less than symbols from the listing. I hope its not too confusing.

00:08.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61) (prog-if 00 [UHCI])
Subsystem: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx-Latency: 128, Cache Line Size: 32 bytes
Interrupt: pin A routed to IRQ 9
Region 4: I/O ports at fe802400 [size=32]
Capabilities: [80] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: uhci_hcd

00:08.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61) (prog-if 00 [UHCI])
Subsystem: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx-
Latency: 128, Cache Line Size: 32 bytes
Interrupt: pin B routed to IRQ 10
Region 4: I/O ports at fe802420 [size=32]
Capabilities: [80] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: uhci_hcd

00:08.2 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 63) (prog-if 20 [EHCI])
Subsystem: VIA Technologies, Inc. USB 2.0
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx-Latency: 128, Cache Line Size: 32 bytes
Interrupt: pin C routed to IRQ 11
Region 0: Memory at 88100100 (32-bit, non-prefetchable) [size=256]
Capabilities: [80] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: ehci_hcd

00:08.3 RAID bus controller: VIA Technologies, Inc. VT6421 IDE RAID Controller (rev 50)
Subsystem: VIA Technologies, Inc. VT6421 IDE RAID Controller
Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx-
Latency: 128
Interrupt: pin A routed to IRQ 9
Region 0: I/O ports at fe802440 [size=16]
Region 1: I/O ports at fe802450 [size=16]
Region 2: I/O ports at fe802460 [size=16]
Region 3: I/O ports at fe802470 [size=16]
Region 4: I/O ports at fe802480 [size=32]
Region 5: I/O ports at fe802500 [size=256]
Expansion ROM at [disabled]
Capabilities: [e0] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: sata_via
Kernel modules: sata_via

Last edited by scabit on 29-Nov-2012 at 08:54 PM.
Last edited by scabit on 29-Nov-2012 at 08:52 PM.
Last edited by scabit on 29-Nov-2012 at 08:49 PM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Quote:

scabit wrote: @Geri Here is the dmesg stuff relating to my usb card..... libata version 2.00 loaded. ehci_hcd 0000:00:08.2: EHCI Host Controller ehci_hcd 0000:00:08.2: new USB bus registered, assigned bus number 1 ehci_hcd 0000:00:08.2: irq 11, io mem 0x88100100 ehci_hcd 0000:00:08.2: USB 2.0 started, EHCI 1.00, driver 10 Dec 2004 usb usb1: configuration #1 chosen from 1 choice hub 1-0:1.0: USB hub found hub 1-0:1.0: 4 ports detected ohci_hcd: 2005 April 22 USB 1.1 'Open' Host Controller (OHCI) Driver (PCI) USB Universal Host Controller Interface driver v3.0 uhci_hcd 0000:00:07.2: UHCI Host Controller uhci_hcd 0000:00:07.2: new USB bus registered, assigned bus number 2 uhci_hcd 0000:00:07.2: irq 5, io base 0x00802080 usb usb2: configuration #1 chosen from 1 choice hub 2-0:1.0: USB hub found hub 2-0:1.0: 2 ports detected uhci_hcd 0000:00:07.3: UHCI Host Controller uhci_hcd 0000:00:07.3: new USB bus registered, assigned bus number 3 uhci_hcd 0000:00:07.3: irq 5, io base 0x008020a0 usb usb3: configuration #1 chosen from 1 choice hub 3-0:1.0: USB hub found hub 3-0:1.0: 2 ports detected uhci_hcd 0000:00:08.0: UHCI Host Controller uhci_hcd 0000:00:08.0: new USB bus registered, assigned bus number 4 uhci_hcd 0000:00:08.0: irq 9, io base 0x00802400 usb usb4: configuration #1 chosen from 1 choice hub 4-0:1.0: USB hub found hub 4-0:1.0: 2 ports detected uhci_hcd 0000:00:08.1: UHCI Host Controller uhci_hcd 0000:00:08.1: new USB bus registered, assigned bus number 5 uhci_hcd 0000:00:08.1: irq 10, io base 0x00802420 usb usb5: configuration #1 chosen from 1 choice hub 5-0:1.0: USB hub found hub 5-0:1.0: 2 ports detected Initializing USB Mass Storage driver... usbcore: registered new driver usb-storage USB Mass Storage support registered. usbcore: registered new driver hiddev usbcore: registered new driver usbhid drivers/usb/input/hid-core.c: v2.6:USB HID core driver

Looks good. What does lsusb say, if you plug in a USB device?

Quote:

Here is the lspci -vvv for the SYBA card USB and IDE/SATA controller 00:08.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61) (prog-if 00 [UHCI]) Subsystem: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 128, Cache Line Size: 32 bytes Interrupt: pin A routed to IRQ 9 Region 4: I/O ports at fe802400 [size=32] Capabilities: [80] Power Management version 2 Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+) Status: D0 PME-Enable- DSel=0 DScale=0 PME- Kernel driver in use: uhci_hcd 00:08.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61) (prog-if 00 [UHCI]) Subsystem: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 128, Cache Line Size: 32 bytes Interrupt: pin B routed to IRQ 10 Region 4: I/O ports at fe802420 [size=32] Capabilities: [80] Power Management version 2 Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+) Status: D0 PME-Enable- DSel=0 DScale=0 PME- Kernel driver in use: uhci_hcd 00:08.2 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 63) (prog-if 20 [EHCI]) Subsystem: VIA Technologies, Inc. USB 2.0 Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 128, Cache Line Size: 32 bytes Interrupt: pin C routed to IRQ 11 Region 0: Memory at 88100100 (32-bit, non-prefetchable) [size=256] Capabilities: [80] Power Management version 2 Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=375mA PME(D0+,D1+,D2+,D3hot+,D3cold+) Status: D0 PME-Enable- DSel=0 DScale=0 PME- Kernel driver in use: ehci_hcd 00:08.3 RAID bus controller: VIA Technologies, Inc. VT6421 IDE RAID Controller (rev 50) Subsystem: VIA Technologies, Inc. VT6421 IDE RAID Controller Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 128 Interrupt: pin A routed to IRQ 9 Region 0: I/O ports at fe802440 [size=16] Region 1: I/O ports at fe802450 [size=16] Region 2: I/O ports at fe802460 [size=16] Region 3: I/O ports at fe802470 [size=16] Region 4: I/O ports at fe802480 [size=32] Region 5: I/O ports at fe802500 [size=256] Expansion ROM at [disabled] Capabilities: [e0] Power Management version 2 Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-) Status: D0 PME-Enable- DSel=0 DScale=0 PME- Kernel driver in use: sata_via Kernel modules: sata_via

Nothing eye-catching. But it could also be that the Expansion ROM needs to initialize the controller chip. Is this card known to work on anything else than the x86 architecture?

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Hypex

Quote:

Hypex wrote: @Geri Quote: It just mounts the partition automatically during boot. What in fstab? Or simply havving an AFFS volume wlil befound and mounted at boot by Linux?

Yes, if there is a (correct) entry in fstab, then it will be mounted automatically.


Quote:

Hypex wrote: @Geri Quote: What do you mean with "looses the drive controller"? What I mean is it doesn't detect I'm only using the 0680 IDE and starts looking for drives on the VIA IDE. So it spends its time looking for drives and resetting the busses only to fail and drop into the initramfs.

You could blacklist the via82cxxx IDE driver and regenerate the initramfs to exclude the driver from it (that's the good thing about initramfs ). Then the Linux kernel will not probe the VIA controller at all.

Quote:

Of course this means I can't rely directly on specifying a drive as /dev/hdxy anywhere. Since my drive could be on /dev/hdc or /dev/hdg.

That shouldn't be the case, if you blacklist the driver.

Quote:

I wish the kennel would look at the UBoot vars!

Uh yeah, that's on the TODO list since several years.

Quote:

But I'd rather a controllerskip option.

Quote:

One thing I also notice. I also can get lots of USB messages. Unable to add hub device or something. I then plug in a drive onto my front USB port and it shuts up! :?

Does your A1 have the USB fix?

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri

I do not have time to run lsusb - immediately when I plug anything into the SYBA USB port themachine locks up. Physical reset required.

As far as if this card works iwth anything other than x86...good question. I cannot find anything online about that. I can ask the manufacturer and see if they respond.
But - as for the USB - this card works as a USB 1.1 controller in OS4.1 IF I disable the EHCI module from kicklayout. It too locks up when I try to use it as USB2 in AmigaOS4.1 u5.

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Quote:

scabit wrote: @Geri I do not have time to run lsusb - immediately when I plug anything into the SYBA USB port themachine locks up. Physical reset required.

Sorry, forgot about that.

Quote:

As far as if this card works iwth anything other than x86...good question. I cannot find anything online about that. I can ask the manufacturer and see if they respond. But - as for the USB - this card works as a USB 1.1 controller in OS4.1 IF I disable the EHCI module from kicklayout. It too locks up when I try to use it as USB2 in AmigaOS4.1 u5.

Well, the USB controller could also trigger another bug in the northbridge, which as far as I understand it sometimes can lock up the system on a PCI MMIO read access, if the CPU tries to access the MMIO space at the same time. I still puzzle how to debug such a lock up, because I think it also happens with the radeon DRM driver.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri


Quote:

You mean the ALSA patch on amigaone-linux.sourceforge.net?

No, there is a specific patch for EHCI USB with this comment:

( 29/06/2008 ) Patch written by Marcus Comstedt. Makes the USB EHCI driver work on non cache coherent PPC 74xx systems

...also on sourceforge in the A1 linux download page.

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@Geri

Quote:

Yes, if there is a (correct) entry in fstab, then it will be mounted automatically.

After manually adding it in.

Quote:

You could blacklist the via82cxxx IDE driver and regenerate the initramfs

I could and I know we've been over this but it would be nice to specify some boot args to disable the onboard IDE. Or even have the kernel read the UBoot vars to see what hardware is disabled.

For a new installer this is really needed if it installs a bootable menu. Gets a bit much asking the user to peform an operation on the initramfs after Linux was just installed and everything setup to boot automatically.

Quote:

That shouldn't be the case, if you blacklist the driver.

Can it be blacklisted in /etc? The kernal command line could be read and a /etc modified to suit before boot.

Forget that it's a kernel issue.

Quote:

TODO l

Let's do it now!

Quote:

Does your A1 have the USB fix?

Not the proper one. But I did ""hack" a couple of resistors onto the front ports.

See here:
http://amigaworld.net/modules/features/index.php?op=r&cat_id=4&rev_id=169&sort_by

I had to find it by hand!

@scabit

Quote:

Makes the USB EHCI driver work on non cache coherent PPC 74xx systems

Wonder if this would help get my VIA USB2 card working in OS4? Currently only USB 1.0/1.1 works. EHCI is broken. Device 0x3104. Revision: 0x65. Vendor: 0x1106.

@Hypex

Quote:

Wonder if this would help get my VIA USB2 card working in OS4? Currently only USB 1.0/1.1 works. EHCI is broken. Device 0x3104. Revision: 0x65. Vendor: 0x1106.

Ahhh..that is the point of my exercise! I have the same VIA USB2 chip in my SYBA card that you have, same revision too I think. Anyway, I suspect the issue with OS4.1 is this dma cache coherency problem with EHCI. If I can prove the card works in linux on my hardware with USB2, then I can start looking at how we can implement a fix for the OS4.1 EHCI kernel module. But if it never works in linux - even with these cache coherency problem fixes, then we'll just have to acknowledge that it is probably a hardware problem that cannot easily be solved.

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@Geri

Quote:

But it could also be that the Expansion ROM needs to initialize the controller chip. Is this card known to work on anything else than the x86 architecture?

Here is the answer to that.....



RE: Does the SYBA SD-VIA-1A5E1lR card SATA ROM work with a PowerPC CPU?

Hi Scott,

Thank you for your reply.

We apologize, but No it Does Not.


Thank you.

On 11/30/2012 4:15 AM, Scott Cabit wrote:
Hi,

The part number of the SYBA card is: SD-VIA-1A5E1IR.
The computer it will be used in can be found here:

http://amigakit.leamancomputing.com/catalog/product_info.php?products_id=1040

The question is, does the SYBA SD-VIA-1A5E1lR ROM for onboard SATA support PowerPC CPUs?

Thank you!

Last edited by scabit on 01-Dec-2012 at 02:19 AM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

That's a start. Now can you ask why?

@Hypex

Quote:

That's a start. Now can you ask why?

It was difficult enough for them to understand what I was asking the first time! It took a couple of email messages back and forth. I got this card from China.

As for the ROM...do any devices really still use "burned" ROMs, or is this ROM really a flash device that can be reflashed? Since there seem to be other cards out there that use this IDE/SATA combo chip and (I think) work with PowerPC processors, it would be a matter of obtaining the code necessary to initialize this chip and reflash the "ROM" if it can be done.
Why is the easy stuff always so difficult?

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@Hypex

Quote:

Hypex wrote: @Geri Quote: You could blacklist the via82cxxx IDE driver and regenerate the initramfs I could and I know we've been over this but it would be nice to specify some boot args to disable the onboard IDE. Or even have the kernel read the UBoot vars to see what hardware is disabled.

Well, Linux supports a different way of ignoring disabled IDE ports. On x86 the BIOS sets the progif register of the IDE controller, which specifies the available IDE channels. The right way would be to add this functionality also to U-Boot.

Quote:

For a new installer this is really needed if it installs a bootable menu. Gets a bit much asking the user to peform an operation on the initramfs after Linux was just installed and everything setup to boot automatically.

That's true. Maybe I can hack something into the platform code to clear the progif register before the driver is loaded.

Quote:

Can it be blacklisted in /etc? The kernal command line could be read and a /etc modified to suit before boot. Forget that it's a kernel issue.

Well, blacklists for kernel drivers are usually setup in /etc/modprobe.d. Would you do this for the installer or for a freshly installed system?

Quote:

Quote: TODO l Let's do it now!

The NVRAM drivers usually only provide a block interface for userspace access, not for the kernel.

Quote:

Quote: Does your A1 have the USB fix? Not the proper one. But I did ""hack" a couple of resistors onto the front ports. See here: http://amigaworld.net/modules/features/index.php?op=r&cat_id=4&rev_id=169&sort_by I had to find it by hand!

I think there were also some capacitors to be removed. But I don't know, if that makes a different. Can you post an excerpt of the USB messages?

Quote:

Hypex wrote: @scabit Quote: Makes the USB EHCI driver work on non cache coherent PPC 74xx systems Wonder if this would help get my VIA USB2 card working in OS4? Currently only USB 1.0/1.1 works. EHCI is broken. Device 0x3104. Revision: 0x65. Vendor: 0x1106.

I really doubt that, because it fixes a Linux specific driver issue.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@scabit

Quote:

scabit wrote: @Hypex Quote: That's a start. Now can you ask why? It was difficult enough for them to understand what I was asking the first time! It took a couple of email messages back and forth. I got this card from China. As for the ROM...do any devices really still use "burned" ROMs, or is this ROM really a flash device that can be reflashed? Since there seem to be other cards out there that use this IDE/SATA combo chip and (I think) work with PowerPC processors, it would be a matter of obtaining the code necessary to initialize this chip and reflash the "ROM" if it can be done. Why is the easy stuff always so difficult?

Well, usually ROMs on cheap IDE/SATA controllers only implement a software RAID functionality, which we are not interested in. I still suspect that the controller chip can't handle PCI I/O addresses outside of a 64kB address space. Maybe we can hack the PCI setup with some U-Boot commands to confirm this theory.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@scabit

Can you please post the content of /proc/ioports and /proc/iomem?

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@scabit

Quote:

It was difficult enough for them to understand what I was asking the first time!

Yeah I once asked HP for technical information about their PCL commands. The end result is them asking me more questions about why I need that information! And usually not giving it but replying back something else useless.

Quote:

As for the ROM...do any devices really still use "burned" ROMs, or is this ROM really a flash device that can be reflashed?

It's possible it is flashable if it can be updated. But I don' think this would help. The code needs to be executed regardless. It could be like the old SCSI PC cards that had their own SCSI BIOS. Infact a few SCSI cards that should have been OS4 compatible didn't work or at least didn't show any devices.

So we support x86 BIOS on VGA cards but nothing else AFAIK. UBoot would need to scan the PCI bus for device ROMs and execute the x86 code on the chip.

Quote:

Why is the easy stuff always so difficult?

Annoying I know! Perhaps our hardware just makes it harder for us to use.

But at this point I wonder if the Linux kernel can scan the PCI list and see if any devices have ROMs. And if so execute the code in there on an x86 emulator. Perhaps it would be best if the driver did this. The driver may assume the code was already ran at bootup. Then all we would need is for the driver to execute the reset code when it launches. If that is the case.

@Geri

Quote:

The right way would be to add this functionality also to U-Boot.

That's interesting. the closest we have is the bootloader. But a delay caused by an empty controller isn't a big deal if it gets past it.

What's more important is being able to set DMA. Needed for install as well as installed system.

Quote:

Maybe I can hack something into the platform code to clear the progif register before the driver is loaded.

Possible but setting DMA is easier?

Quote:

Would you do this for the installer or for a freshly installed system?

Quote:

o clear the progif register before the driver is loaded.Quote:

Fresh install would need it. But it also needs to install without errors!

Quote:

The NVRAM drivers usually only provide a block interface for userspace access, not for the kernel.

Okay if we can read it and deal with it on time.

Quote:

Can you post an excerpt of the USB messages?

Sure. Because my fix is a solderless one the capacitors stay on!

It starts out:
[ 38.781915] hub 3-0:1.0: USB hub found
[ 38.808774] hub 3-0:1.0: 2 ports detected
[ 38.810668] uhci_hcd 0000:00:09.0: calling quirk_via_vlink+0x0/0xc4
[ 38.810712] uhci_hcd 0000:00:09.0: UHCI Host Controller
[ 38.836480] uhci_hcd 0000:00:09.0: new USB bus registered, assigned bus number 4
[ 38.837747] uhci_hcd 0000:00:09.0: irq 10, io base 0x00802440
[ 38.847353] usb usb4: New USB device found, idVendor=1d6b, idProduct=0001
[ 38.848454] usb usb4: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[ 38.849446] usb usb4: Product: UHCI Host Controller
[ 38.850393] usb usb4: Manufacturer: Linux 2.6.38.3+ uhci_hcd
[ 38.851396] usb usb4: SerialNumber: 0000:00:09.0

If usb4 corresponds to hub 3. Log doesn't make it clear to me. But this kind of message constantly appears every second or so as it boots:

[ 41.762992] hub 3-0:1.0: unable to enumerate USB device on port 2

Quote:

I really doubt that, because it fixes a Linux specific driver issue.

Then I'll guess I'll never know what OS4 is doing wrong.