Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.

Dismiss

Cyclades-8Y on G4(Yikes!)

3 views

Skip to first unread message

Bill Studenmund

unread,

Jul 25, 2000, 3:00:00 AM7/25/00

On Thu, 13 Jul 2000, Andreas Wrede wrote:

> I can successful netboot the 20000620-1.5 snapshot on a G4/400
> machine - dmesg output below.

Did you get any help with this?

I know it might not help much, but the cz serial cards work just fine. I
debugged a Cyclades-Z 16e on my G3. :-)

> The Cylcades-8Y PCI card does get recognized, and after making the
> /dev nodes manually (no support in MAKEDEV, send-pr to follow), I can
> open a port on the card. The modem signals appear to be handled
> correctly, ie. the attached modem lights DTR when the port is opened,
> and kermit reports correct status of DCD and DSR. However, I cannot
> send or receive any characters through the port. systat vmstat does
> not list the irq (irq24) in the right column, which leads me to
> believe that something is wrong with the interrupt management in the
> driver.
>
> Is someone using a PCI Cyclades-Y board on G3 or G4 Powermac with
> NetBSD 1.4/5? If not, where should I start to look in the code to
> trace the missing interrupts?

syssrc/sys/dec/pci/cy_pci.c is the pci-specific code for the driver, and
syssrc/sys/dev/ic/cy.c is the bus-independent part of the driver.

> $ ls -l /dev/*tyCY00
> crw------- 1 uucp wheel 47, 128 Jul 13 11:50 /dev/dtyCY00
> crw------- 1 uucp wheel 47, 0 Jul 13 10:26 /dev/ttyCY00

You might also talk with tsubai, as he was the one who enabled cy
support. :-)

Take care,

Bill

Andreas Wrede

unread,

Jul 25, 2000, 3:00:00 AM7/25/00

On Mon, 24 Jul 2000, Bill Studenmund wrote:

> On Thu, 13 Jul 2000, Andreas Wrede wrote:
>
> > I can successful netboot the 20000620-1.5 snapshot on a G4/400
> > machine - dmesg output below.
>
> Did you get any help with this?

Not yet - I am still hoping to get this to work.

> I know it might not help much, but the cz serial cards work just fine. I
> debugged a Cyclades-Z 16e on my G3. :-)

:-) I suspect the driver will work just fine once the interrupts get
delivered from the card - I have never seem a single one. The card
currently is an B&W G3, due to the problem with the PCI0646 ide
controller chip initialization after the G4 firmware has had its hands
on it. The Cyclom8Y has the same problem in both the G4 and G3 - no
interrupts. Modem signals are delivered properly.

> > The Cylcades-8Y PCI card does get recognized, and after making the
> > /dev nodes manually (no support in MAKEDEV, send-pr to follow), I can
> > open a port on the card. The modem signals appear to be handled
> > correctly, ie. the attached modem lights DTR when the port is opened,
> > and kermit reports correct status of DCD and DSR. However, I cannot
> > send or receive any characters through the port. systat vmstat does
> > not list the irq (irq24) in the right column, which leads me to
> > believe that something is wrong with the interrupt management in the
> > driver.
> >
> > Is someone using a PCI Cyclades-Y board on G3 or G4 Powermac with
> > NetBSD 1.4/5? If not, where should I start to look in the code to
> > trace the missing interrupts?
>
> syssrc/sys/dec/pci/cy_pci.c is the pci-specific code for the driver, and
> syssrc/sys/dev/ic/cy.c is the bus-independent part of the driver.

I found these two. Where does PCI bus interrupt routing happen? Or is
that to naive a question to ask?

>
> > $ ls -l /dev/*tyCY00
> > crw------- 1 uucp wheel 47, 128 Jul 13 11:50 /dev/dtyCY00
> > crw------- 1 uucp wheel 47, 0 Jul 13 10:26 /dev/ttyCY00
>
> You might also talk with tsubai, as he was the one who enabled cy
> support. :-)

I Cc:'d tsubai, maybe he is not on port-macppc anymore?

I compiled the cy driver with CY_DEBUG on. Can anyone find a clue as
to why interrupts are not working at all??

pciide0: secondary channel ignored (disabled)
cy: Found Cyclades PCI device, id = 0x105120e
cy0 at pci1 dev 4 function 0cy: card reset done
cy0 probe chip 0 offset 0x0 ... firmware version 0x48
cy0 probe chip 1 offset 0x800 ... firmware version 0x48
cy0 probe chip 2 offset 0x1000 ... not ready for command
found 2 CD1400s
: interrupting at irq 25
cy0attach CD1400 #0 offset 0x0
attach CD1400 #1 offset 0x800
: 8 ports
obio0 at pci1 dev 5 function 0: addr 0x80800000
[...]

root on wd0a dumps on wd0b
root file system type: ffs
cy0 open port 0 flag 0x5 mode 0x2000
c1400_channel_cmd cy 0xe0221060 command 0x80
cy0: port 0 param tty 0xe00378 termios 0xe00450
ispeed 9600 ospeed 9600
c1400_channel_cmd cy 0xe0221060 command 0x1a
cor1 = 0x13...c1400_channel_cmd cy 0xe0221060 command 0x4e
cy0: port 0 ioctl cmd 0x8004745c data 0xe975fe28 flag 0x5
cy0: close port 0, flag 0x5, mode 0x2000
cy0: port 0 stop tty 0xe00378 flag 0x3
cy0: port 0 stop tty 0xe00378 flag 0x3
cy12033 open port 1 flag 0x5 mode 0x2000
c1400_channel_cmd cy 0xe02210a0 command 0x80
cy0: port 1 param tty 0xe004a0 termios 0xe00578
ispeed 9600 ospeed 9600
c1400_channel_cmd cy 0xe02210a0 command 0x1a
cor1 = 0x13...c1400_channel_cmd cy 0xe02210a0 command 0x4e
cy0: port 1 ioctl cmd 0x8004745c data 0xe975fe28 flag 0x5
cy0: close port 1, flag 0x5, mode 0x2000
cy0: port 1 stop tty 0xe004a0 flag 0x3
cy0: port 1 stop tty 0xe004a0 flag 0x3

--
- aew

Tsubai Masanari

unread,

Jul 25, 2000, 3:00:00 AM7/25/00

>The Cyclom8Y has the same problem in both the G4 and G3 - no
>interrupts. Modem signals are delivered properly.

At least, on my 9600 it is working. Probably uninorth is wrong?

Cliff Neighbors

unread,

Jul 25, 2000, 3:00:00 AM7/25/00

This message is in MIME format. Since your mail reader does not understand
this format, some or all of this message may not be legible.

------_=_NextPart_001_01BFF5EC.6568B0B0
Content-Type: text/plain;
charset="iso-8859-1"

possibly premature to say, but smells like similar
symptoms to what I am running into with the SIIG 4S
16550 card and the com_puc driver: card gets
detected at boot up OK but never sees an interrupt.
This on a 400MHz G4 running a kernel based on the
20000606 snapshot.

-cliff-

---
cliff neighbors
allegro networks, inc.
cl...@allegronetworks.com
408-821-5532
---

------_=_NextPart_001_01BFF5EC.6568B0B0
Content-Type: text/html;
charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2652.35">
<TITLE>RE: Cyclades-8Y on G4(Yikes!)</TITLE>
</HEAD>
<BODY>

possibly premature to say, but smells like similar
 symptoms to what I am running into with the SIIG 4S
 16550 card and the com_puc driver: card gets
 detected at boot up OK but never sees an interrupt.
 This on a 400MHz G4 running a kernel based on the 
 20000606 snapshot.

-cliff-

---
 cliff neighbors
 allegro networks, inc.
 cl...@allegronetworks.com
 408-821-5532
 ---

-----Original Message-----
 From: Tsubai Masanari [<A HREF="mailto:tsu...@iri.co.jp">mailto:tsu...@iri.co.jp</A>]
 Sent: Monday, July 24, 2000 7:53 PM
 To: Andreas Wrede
 Cc: port-...@netbsd.org
 Subject: Re: Cyclades-8Y on G4(Yikes!)

>The Cyclom8Y has the same problem in both the G4 and G3 - no
 >interrupts. Modem signals are delivered properly.

At least, on my 9600 it is working.  Probably uninorth is wrong?

</BODY>
</HTML>
------_=_NextPart_001_01BFF5EC.6568B0B0--

Bill Studenmund

unread,

Jul 25, 2000, 3:00:00 AM7/25/00

On Mon, 24 Jul 2000, Andreas Wrede wrote:

> On Mon, 24 Jul 2000, Bill Studenmund wrote:
>
> > I know it might not help much, but the cz serial cards work just fine. I
> > debugged a Cyclades-Z 16e on my G3. :-)
>
> :-) I suspect the driver will work just fine once the interrupts get
> delivered from the card - I have never seem a single one. The card
> currently is an B&W G3, due to the problem with the PCI0646 ide
> controller chip initialization after the G4 firmware has had its hands

> on it. The Cyclom8Y has the same problem in both the G4 and G3 - no

> interrupts. Modem signals are delivered properly.

I think the B&W G3 has the same ide problem. I'll assume that you just
also have a SCSI card too. :-)

> > > Is someone using a PCI Cyclades-Y board on G3 or G4 Powermac with
> > > NetBSD 1.4/5? If not, where should I start to look in the code to
> > > trace the missing interrupts?
> >
> > syssrc/sys/dec/pci/cy_pci.c is the pci-specific code for the driver, and
> > syssrc/sys/dev/ic/cy.c is the bus-independent part of the driver.
>
> I found these two. Where does PCI bus interrupt routing happen? Or is
> that to naive a question to ask?

I'm not sure. Though I'd not suspect interrupt routing problems at
first. To test, get another PCI card which will interrupt, and put it in
the slot. make sure it says it's using the same interrupt # as the cy card
(PCI cards have 4 interrupts each). If it works, you have verified the
interrupt and routing is fine.

> > > $ ls -l /dev/*tyCY00
> > > crw------- 1 uucp wheel 47, 128 Jul 13 11:50 /dev/dtyCY00
> > > crw------- 1 uucp wheel 47, 0 Jul 13 10:26 /dev/ttyCY00
> >
> > You might also talk with tsubai, as he was the one who enabled cy
> > support. :-)
>
> I Cc:'d tsubai, maybe he is not on port-macppc anymore?

He's the portmaster, so I'd surprised if he weren't around. :-)

> I compiled the cy driver with CY_DEBUG on. Can anyone find a clue as
> to why interrupts are not working at all??

No, but then I don't know much about this card. I suspect that the real
problem is that interrupts on the card aren't getting enabled right. I
can't figure out why, though.

Ask on teck-kern. I'm not too familiar with this driver, and I think the
folks who've worked on it will be there. :-)

Take care,

Bill

0 new messages