Lite-On ethernet controller

Lite-On ethernet controller

[Josh Huber]

I have what appears to be a Lite-On ethernet controller, which is supported
by the tulip driver. What it really is, is a Asante FriendlyNET card. I had
heard that the card works in 2.2.14+, but I'm having trouble getting it to
work in later kernels (devel).

Because of the network driver interface change, I can't try out the newest
tulip driver in 2.4pre.  Does anyone know of forward-ported tulip drivers?

The particular error I'm seeing only happens with moderate load -- pings and
other light traffic work 100%.

Error information:
NETDEV WATCHDOG: eth1: transmit timed out
Transmit timed out, status e4660000, CSR12 000050ca, resetting...

Device information:
tulip driver output:
eth1: Lite-On PNIC-II rev 37 at 0x1c00, 00:00:94:C5:EF:FF, IRQ 30.
lspci -vv output attached because it's too wide.

Any ideas?

Josh

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Jeff Garzik]

Josh Huber wrote:
>
> I have what appears to be a Lite-On ethernet controller, which is supported
> by the tulip driver. What it really is, is a Asante FriendlyNET card. I had
> heard that the card works in 2.2.14+, but I'm having trouble getting it to
> work in later kernels (devel).
>
> Because of the network driver interface change, I can't try out the newest
> tulip driver in 2.4pre.

I don't understand what you mean here.


> Does anyone know of forward-ported tulip drivers?

A tulip driver exists in the 2.4pre kernel.  Are you having problems
with that driver?


> The particular error I'm seeing only happens with moderate load -- pings and
> other light traffic work 100%.
>
> Error information:
> NETDEV WATCHDOG: eth1: transmit timed out
> Transmit timed out, status e4660000, CSR12 000050ca, resetting...
>
> Device information:
> tulip driver output:
> eth1: Lite-On PNIC-II rev 37 at 0x1c00, 00:00:94:C5:EF:FF, IRQ 30.

IRQ 30?  Is that ok for PPC?


> lspci -vv output attached because it's too wide.

No attachment...

	Jeff



--
Jeff Garzik              | Liberty is always dangerous, but
Building 1024            | it is the safest thing we have.
MandrakeSoft, Inc.       |      -- Harry Emerson Fosdick

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Josh Huber]

[-- Attachment #1: Type: text/plain, Size: 1613 bytes --]

On Wed, May 31, 2000 at 09:32:12AM -0400, Jeff Garzik wrote:
> I don't understand what you mean here.
Sorry I wasn't clear.

The Tulip driver in the devel kernels is not working for me. Supposedly,
this card works with the driver in later 2.2 kernels.  I can't get 2.2.x to
boot on this machine I have, which is why I'm using 2.4 pre versions.

What I meant by the newer tulip driver was the version available from Don
Becker's tulip development page.

> > Does anyone know of forward-ported tulip drivers?
>
> A tulip driver exists in the 2.4pre kernel.  Are you having problems
> with that driver?
Yes.

> > The particular error I'm seeing only happens with moderate load -- pings and
> > other light traffic work 100%.
> >
> > Error information:
> > NETDEV WATCHDOG: eth1: transmit timed out
> > Transmit timed out, status e4660000, CSR12 000050ca, resetting...
> >
> > Device information:
> > tulip driver output:
> > eth1: Lite-On PNIC-II rev 37 at 0x1c00, 00:00:94:C5:EF:FF, IRQ 30.
>
> IRQ 30?  Is that ok for PPC?
Yes, here's the output from /proc/interrupts:
              CPU0
     1:       1820   i8259         keyboard
     2:          0   i8259         82c59 secondary cascade
     8:          2   i8259         rtc
    16:          0   OpenPIC       82c59 cascade
    27:          0   OpenPIC       NMI
    29:       4872   OpenPIC       eth0
    30:       4624   OpenPIC       eth1
    31:       6377   OpenPIC       aic7xxx

eth0 is an actual DEC tulip card, which works great.

> > lspci -vv output attached because it's too wide.
> No attachment...

Heh, sorry.  I'll attach it this time.

Josh

[-- Attachment #2: lspci.out --]
[-- Type: text/plain, Size: 808 bytes --]

00:05.0 Ethernet controller: Lite-On Communications Inc: Unknown device c115 (rev 25)
        Subsystem: Asante Technologies, Inc.: Unknown device f001
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B-
        Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Latency: 8 min, 56 max, 32 set, cache line size 08
        Interrupt: pin A routed to IRQ 30
        Region 0: I/O ports at 1c00 [size=256]
        Region 1: Memory at c3002000 (32-bit, non-prefetchable) [size=256]
        Expansion ROM at c1240000 [disabled] [size=256K]
        Capabilities: [44] Power Management version 1
                Flags: PMEClk- AuxPwr- DSI- D1+ D2+ PME-
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Re: Lite-On ethernet controller

[Michel Lanners]

Hi there,

On  31 May, this message from Josh Huber echoed through cyberspace:
>> > lspci -vv output attached because it's too wide.


00:05.0 Ethernet controller: Lite-On Communications Inc: Unknown device c115 (rev 25)
        Subsystem: Asante Technologies, Inc.: Unknown device f001
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B-
        Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Latency: 8 min, 56 max, 32 set, cache line size 08
        Interrupt: pin A routed to IRQ 30
        Region 0: I/O ports at 1c00 [size=256]
                  ^^^^^^^^^    ^^^^
        Region 1: Memory at c3002000 (32-bit, non-prefetchable) [size=256]
        Expansion ROM at c1240000 [disabled] [size=256K]

The IO port region above might be your problem. What machine type is
this? Do you have any ISA slots in your machine? Form the i8259
interrupt controller I deduce it's not a PowerMac...

Have you looked at the driver whether it runs on IO ports or
memory-mapped? Since the two regions above probably represent different
views on the same register file, I'd go memory-mapped (i.e. using region
1 above), if possible.

Michel

-------------------------------------------------------------------------
Michel Lanners                 |  " Read Philosophy.  Study Art.
23, Rue Paul Henkes            |    Ask Questions.  Make Mistakes.
L-1710 Luxembourg              |
email   mlan@cpu.lu            |
http://www.cpu.lu/~mlan        |                     Learn Always. "


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Josh Huber]

On Wed, May 31, 2000 at 08:40:57PM +0200, Michel Lanners wrote:
> 00:05.0 Ethernet controller: Lite-On Communications Inc: Unknown device c115 (rev 25)
>         Subsystem: Asante Technologies, Inc.: Unknown device f001
>         Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B-
>         Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
>         Latency: 8 min, 56 max, 32 set, cache line size 08
>         Interrupt: pin A routed to IRQ 30
>         Region 0: I/O ports at 1c00 [size=256]
>                   ^^^^^^^^^    ^^^^
>         Region 1: Memory at c3002000 (32-bit, non-prefetchable) [size=256]
>         Expansion ROM at c1240000 [disabled] [size=256K]
>
> The IO port region above might be your problem. What machine type is
> this? Do you have any ISA slots in your machine? Form the i8259
> interrupt controller I deduce it's not a PowerMac...

It's not a powermac...it's a CHRP Longtrail board.  Geert's board has an ISA
slot, and I believe mine has all the hardware for an ISA slot, except the
actual slot was removed.

Geert, how did you get the tulip-diag program to compile on your chrp
machine?  I'd like to give it a try and see what it tells me.

> Have you looked at the driver whether it runs on IO ports or
> memory-mapped? Since the two regions above probably represent different
> views on the same register file, I'd go memory-mapped (i.e. using region
> 1 above), if possible.

How would I check this?  Perhaps Jeff knows this? :)

What's the next step for debugging this?

Now, as for possible buggy motherboards, wouldn't other PCI devices have
trouble?  Perhaps network cards just show the problem becuase of the high
interrupt delivery rate?

I've got other PCI cards in this machine, and haven't had any trouble with
them:
00:00.0 Host bridge: VLSI Technology Inc VAS96011 [Golden Gate II] (rev 14)
00:01.0 ISA bridge: Symphony Labs W83C553 (rev 05)
00:01.1 IDE interface: Symphony Labs SL82c105 (rev 05)
00:02.0 Class ff00: Apple Computer Inc. Hydra Mac I/O (rev 03)
00:03.0 VGA compatible controller: ATI Technologies Inc 3D Rage I/II 215GT [Mach64 GT] (rev 9a)
00:04.0 Ethernet controller: Digital Equipment Corporation DECchip 21142/43 (rev 30)
00:05.0 Ethernet controller: Lite-On Communications Inc: Unknown device c115 (rev 25)
00:06.0 SCSI storage controller: Adaptec AIC-7881U

Josh

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Geert Uytterhoeven]

On Thu, 1 Jun 2000, Josh Huber wrote:
> On Wed, May 31, 2000 at 08:40:57PM +0200, Michel Lanners wrote:
> > 00:05.0 Ethernet controller: Lite-On Communications Inc: Unknown device c115 (rev 25)
> >         Subsystem: Asante Technologies, Inc.: Unknown device f001
> >         Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B-
> >         Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
> >         Latency: 8 min, 56 max, 32 set, cache line size 08
> >         Interrupt: pin A routed to IRQ 30
> >         Region 0: I/O ports at 1c00 [size=256]
> >                   ^^^^^^^^^    ^^^^
> >         Region 1: Memory at c3002000 (32-bit, non-prefetchable) [size=256]
> >         Expansion ROM at c1240000 [disabled] [size=256K]
> >
> > The IO port region above might be your problem. What machine type is

I/O port is OK, since the card does work most of the time.

> > this? Do you have any ISA slots in your machine? Form the i8259
> > interrupt controller I deduce it's not a PowerMac...
>
> It's not a powermac...it's a CHRP Longtrail board.  Geert's board has an ISA
> slot, and I believe mine has all the hardware for an ISA slot, except the
> actual slot was removed.

CHRP LongTrail has ISA. Interrupts are controlled by an OpenPIC master, and the
traditional dual i8259 cascade as a slave.

> Geert, how did you get the tulip-diag program to compile on your chrp
> machine?  I'd like to give it a try and see what it tells me.

By pathing it to map ISA I/O space :-) I mailed the patch to you.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Geert Uytterhoeven]

On Wed, 31 May 2000, Josh Huber wrote:
> On Wed, May 31, 2000 at 09:32:12AM -0400, Jeff Garzik wrote:
> > I don't understand what you mean here.
> Sorry I wasn't clear.
>
> The Tulip driver in the devel kernels is not working for me. Supposedly,
> this card works with the driver in later 2.2 kernels.  I can't get 2.2.x to
> boot on this machine I have, which is why I'm using 2.4 pre versions.
>
> What I meant by the newer tulip driver was the version available from Don
> Becker's tulip development page.
>
> > > Does anyone know of forward-ported tulip drivers?
> >
> > A tulip driver exists in the 2.4pre kernel.  Are you having problems
> > with that driver?
> Yes.
>
> > > The particular error I'm seeing only happens with moderate load -- pings and
> > > other light traffic work 100%.
> > >
> > > Error information:
> > > NETDEV WATCHDOG: eth1: transmit timed out
> > > Transmit timed out, status e4660000, CSR12 000050ca, resetting...
> > >
> > > Device information:
> > > tulip driver output:
> > > eth1: Lite-On PNIC-II rev 37 at 0x1c00, 00:00:94:C5:EF:FF, IRQ 30.
> >
> > IRQ 30?  Is that ok for PPC?
> Yes, here's the output from /proc/interrupts:
>               CPU0
>      1:       1820   i8259         keyboard
>      2:          0   i8259         82c59 secondary cascade
>      8:          2   i8259         rtc
>     16:          0   OpenPIC       82c59 cascade
>     27:          0   OpenPIC       NMI
>     29:       4872   OpenPIC       eth0
>     30:       4624   OpenPIC       eth1
>     31:       6377   OpenPIC       aic7xxx
>
> eth0 is an actual DEC tulip card, which works great.
>
> > > lspci -vv output attached because it's too wide.
> > No attachment...
>
> Heh, sorry.  I'll attach it this time.

Josh is not the only one having problems...

I have a Digital DC21041 Tulip rev 33 at 0x1080, 21041 mode, 00:80:C8:5A:F8:5B,
IRQ 29. The interrupt number is correct :-) I occasionally got random network
lock ups with the old dex45 driver (tulip didn't work then) that could be
solved by rebooting only (ifconfig down locked up the machine).

Later the card seemed to work fine with the updated tulip driver, as you
probably remember.  However, I just got a problem with the updated driver
(2.3.99-pre3) for the first time:

| NETDEV WATCHDOG: eth0: transmit timed out
| eth0: 21041 transmit timed out, status fc660010, CSR12 000002c8, CSR13 ffffef0d, CSR14 fffff73d, resetting...
| eth0: 21143 100baseTx sensed media.
        ^^^^^
	Huh? I have a real 21041.

So I ran `tulip-diag -fa' and got:

| tulip-diag.c:v1.19 10/2/99 Donald Becker (becker@cesdis.gsfc.nasa.gov)
| Index #1: Found a Digital DC21041 Tulip adapter at 0x1080.
| Digital DC21041 Tulip chip registers at 0x1080:
|   ffe08000 ffffffff ffffffff 07917000 07917200 fc660010 fffe2002 ffffebef
|   fffe0000 ffff4bf8 ffffffff fffe0000 000000c8 ffffef0d fffff73d ffff0006
|  Port selection is half-duplex.
|  Transmit started, Receive started, half-duplex.
|   The Rx process state is 'Waiting for packets'.
|   The Tx process state is 'Idle'.
|   The transmit unit is set to store-and-forward.
|   The NWay status register is 000000c8.
|   Internal autonegotiation state is 'Autonegotiation disabled'.

Fortunately `ifconfig down; ifconfig up' solved the problem and made the
network work again. I reran tulip-diag -fa' and got:

| tulip-diag.c:v1.19 10/2/99 Donald Becker (becker@cesdis.gsfc.nasa.gov)
| Index #1: Found a Digital DC21041 Tulip adapter at 0x1080.
| Digital DC21041 Tulip chip registers at 0x1080:
|   ffe08000 ffffffff ffffffff 07917000 07917200 fc660000 fffe2002 ffffebef
|   fffe0000 ffff4bf8 ffffffff fffe0000 000001c8 ffffef05 ffffff3f ffff0008
|  Port selection is half-duplex.
|  Transmit started, Receive started, half-duplex.
|   The Rx process state is 'Waiting for packets'.
|   The Tx process state is 'Idle'.
|   The transmit unit is set to store-and-forward.
|   The NWay status register is 000001c8.
|   Internal autonegotiation state is 'Autonegotiation disabled'.

I hope this helps. Thanks for your time!

Note that Josh is using a similar box as me, a CHRP LongTrail. In fact all
people with LongTrails seem to have problems with Tulip chips. So it might be a
motherboard problem.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Josh Huber]

On Wed, May 31, 2000 at 10:10:56PM +0200, Geert Uytterhoeven wrote:
> Later the card seemed to work fine with the updated tulip driver, as you
> probably remember.  However, I just got a problem with the updated driver
> (2.3.99-pre3) for the first time:
[snip]

After updating to the newer tulip driver in the devel kernels, my problems
with my real tulip card went away.  I've been having trouble getting two
other cards working however -- the lite-on card that I mentioned before, and
another tulip clone (Kingston).  The kingston card was much worse -- it
dropped ~30% of packets under light load, and completely locked up under
medium-heavy load.

> Note that Josh is using a similar box as me, a CHRP LongTrail. In fact all
> people with LongTrails seem to have problems with Tulip chips. So it might be a
> motherboard problem.
Yeah, which would be unfortunate... :(

Josh

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Jeff Garzik]

Josh Huber wrote:
>
> On Wed, May 31, 2000 at 10:10:56PM +0200, Geert Uytterhoeven wrote:
> > Later the card seemed to work fine with the updated tulip driver, as you
> > probably remember.  However, I just got a problem with the updated driver
> > (2.3.99-pre3) for the first time:
> [snip]
>
> After updating to the newer tulip driver in the devel kernels, my problems
> with my real tulip card went away.  I've been having trouble getting two
> other cards working however -- the lite-on card that I mentioned before, and
> another tulip clone (Kingston).  The kingston card was much worse -- it
> dropped ~30% of packets under light load, and completely locked up under
> medium-heavy load.
>
> > Note that Josh is using a similar box as me, a CHRP LongTrail. In fact all
> > people with LongTrails seem to have problems with Tulip chips. So it might be a
> > motherboard problem.
> Yeah, which would be unfortunate... :(

That would REALLY suck because I was looking forward to throwing my PNIC
and PNIC II test cards into this old PMac 7200 which I use for testing
drivers on PPC.  If it is a general endian bug in the PNIC* codepath,
testing there should hopefully expose the problem.

If not, I have a feeling Josh and I will be exchanging e-mail... :)

	Jeff



--
Jeff Garzik              | Liberty is always dangerous, but
Building 1024            | it is the safest thing we have.
MandrakeSoft, Inc.       |      -- Harry Emerson Fosdick

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Lite-On ethernet controller

[Michel Lanners]

On  31 May, this message from Jeff Garzik echoed through cyberspace:
>> Device information:
>> tulip driver output:
>> eth1: Lite-On PNIC-II rev 37 at 0x1c00, 00:00:94:C5:EF:FF, IRQ 30.
>
> IRQ 30?  Is that ok for PPC?

Yes, actually there are systems that are not limited to 8 IRQs cascaded
into another 8 IRQs ;-))

Seriously, IIRC, desktop PPC systems have at least one interrupt
controller with 32 IRQs. Some even have more... or more controllers.

Michel

-------------------------------------------------------------------------
Michel Lanners                 |  " Read Philosophy.  Study Art.
23, Rue Paul Henkes            |    Ask Questions.  Make Mistakes.
L-1710 Luxembourg              |
email   mlan@cpu.lu            |
http://www.cpu.lu/~mlan        |                     Learn Always. "


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/