NetBSD on Sun's 64-bit Ultrasparc based machines (NetBSD/sparc64)

Michael Thompson | 28 Jul 2011 02:28

Picon

Keyboard Problem with V5.1 on an E4000

I am trying to revive an E4000. I did a "boot cdrom" at the OpenBoot
prompt, NetBSD boots, displays a bunch of configuration messages on
the console, changes the configuration of the video display (it says
wsdisplay0) displays the configuration messages, and asks for the
terminal type. At this point the keyboard stops working.  A "stop-A"
will not get it back to OpenBoot. They keyboard works OK with
OpenSolaris.

Ideas?

--

-- 
Michael Thompson

Martin Husemann | 28 Jul 2011 09:36

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Wed, Jul 27, 2011 at 08:28:48PM -0400, Michael Thompson wrote:
> Ideas?

From the kernel messages, can you tell what variant of {e,u,o}hci the usb
is on and what it says about the keyboard?

Can you try a current snapshot?

Martin

doomwarrior | 28 Jul 2011 20:16

Picon

Re: Keyboard Problem with V5.1 on an E4000


> > From the kernel messages, can you tell what variant of {e,u,o}hci the usb
> is on and what it says about the keyboard?
>
> Can you try a current snapshot?
>
> Martin
>
It is a E4000 - big Iron - this shouldn't have a USB-controller at all!
Normal Keyboard-connector was still type5 keyboard back then - At least 
my E4500 has one.
The E4000 and E4500 are very similar so I doubt there is a difference.

Best Regards

Michael Thompson | 29 Jul 2011 02:19

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Thu, Jul 28, 2011 at 2:16 PM, doomwarrior <doomwarriorx <at> gmail.com> wrote:
>
>> > From the kernel messages, can you tell what variant of {e,u,o}hci the
>> > usb
>> is on and what it says about the keyboard?
>>
>> Can you try a current snapshot?
>>
>> Martin
>>
> It is a E4000 - big Iron - this shouldn't have a USB-controller at all!
> Normal Keyboard-connector was still type5 keyboard back then - At least my
> E4500 has one.
> The E4000 and E4500 are very similar so I doubt there is a difference.
>
> Best Regards
>
>
>

It has the normal Sun non-USB keyboard port on an I/O board in slot 0.
It has a CGSix Sbus board on the I/O board in slot 0.

I pulled all of the CPU boards but one to simplify the configuration.

With the keyboard unplugged and using the serial console:

ok boot cdrom
Boot device: /sbus <at> 3,0/SUNW,fas <at> 3,8800000/sd <at> 6,0:f  File and args:
NetBSD IEEE 1275 Bootblock

(Continue reading)

matthew green | 29 Jul 2011 04:27

Picon

re: Keyboard Problem with V5.1 on an E4000


> avail memory = 981 MB
> mainbus0 (root): SUNW,Ultra-Enterprise (8-slot Sun Enterprise
> E4500/E5500): hostid 809db71e
> cpu0 at mainbus0: SUNW,UltraSPARC-II  <at>  336 MHz, UPA id 0
> cpu0: 16K instruction (32 b/l), 16K data (32 b/l), 4096K external (64 b/l)
> cpu at mainbus0 not configured
> central at mainbus0 not configured
> fhc at mainbus0 not configured

i think one of these is the problem.  either central or fhc should
be attaching something here, and we're missing some of the onboard
devices attached here.  both serial ports and keyboard are missing
from these boot logs.

the reason it works with serial console is that pcons attaches.  the
reason it doesn't work with the keyboard attached is that pcons does
not attach here, presumably because wsdiplay0 claimed the console?

can you send or link to a prtconf -v or whatever it is from solaris
that will show the entire OFW device tree?  it would be great to get
these systems finally supported!

thanks.

.mrg.

matthew green | 29 Jul 2011 10:01

Picon

re: Keyboard Problem with V5.1 on an E4000


so, i ported the fhc and central drivers, and the clock and zs
attachments for them.  however, it crashes when attaching the
central device.  not sure what is up there yet, it crashes in
a very strange way.

myself and others will keep looking.

.mrg.

Julian Coleman | 11 Aug 2011 10:04

Picon

Re: Keyboard Problem with V5.1 on an E4000

Hi,

> so, i ported the fhc and central drivers, and the clock and zs
> attachments for them.  however, it crashes when attaching the
> central device.  not sure what is up there yet, it crashes in
> a very strange way.
> 
> myself and others will keep looking.

Thanks for porting fhc and cnetral.  I tried this out on an E3500 with a
serial console.  Console output is:

  mainbus0 (root): SUNW,Ultra-Enterprise (5-slot Sun Enterprise E3500): hostid 809341f3
  cpu0 at mainbus0: SUNW,UltraSPARC-II  <at>  400 MHz, UPA id 6
  cpu0: 16K instruction (32 b/l), 16K data (32 b/l), 8192K external (64 b/l)
    ...
  cpu7 at mainbus0: SUNW,UltraSPARC-II  <at>  400 MHz, UPA id 19
  cpu7: 16K instruction (32 b/l), 16K data (32 b/l), 8192K external (64 b/l)
  central0 at mainbus0
  fhc0 at central0 board 1: SUNW,501-2511
  clock0 at fhc0: mk48t59
  zs0 at fhc0
  zstty0 at zs0 channel 0 (console i/o)
  zstty1 at zs0 channel 1

  Fatal Reset

   <at> (#) Ultra Enterprise 3.2 Version 30 created 2002/10/25 14:03
  3,0>FATAL ERROR
  3,0>    At time of error: System software was running.

(Continue reading)

Eduardo Horvath | 11 Aug 2011 17:52

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Thu, 11 Aug 2011, Julian Coleman wrote:

> The reset is at the point where we enable interrupts on the serial port (line
> 473 of sys/arch/sparc64/dev/zs.c [1]).  According to the "Sun Enterprise xx00
> Problem Solving Manual", section 4.13 [2], the reset is caused by "an
> interrupt being sent to an unmapped or non-accepting destination".

Looks like the fhc driver is not setting the destination UPAID where it 
needs to when establishing an interrupt mapping.  I don't have the docs on 
the fhc controller, but Sun usually recycled register contents across 
designs.

SBus does this:

                                /* Register the map and clear intr 
registers */
                                ih->ih_map = &intrptr[i];
                                intrptr = (int64_t 
*)&sc->sc_sysio->scsi_clr_int;
                                ih->ih_clr = &intrptr[i];
                                /* Enable the interrupt */
                                imap |= INTMAP_V
                                    |(CPU_UPAID << INTMAP_TID_SHIFT);
                                /* XXXX */
                                *(ih->ih_map) = imap;

FHC does this:

        if (intrmapptr != NULL) {
                u_int64_t r;

(Continue reading)

Julian Coleman | 12 Aug 2011 11:27

Picon

Re: Keyboard Problem with V5.1 on an E4000

Hi,

> Looks like the fhc driver is not setting the destination UPAID where it 
> needs to when establishing an interrupt mapping.  I don't have the docs on 
> the fhc controller, but Sun usually recycled register contents across 
> designs.

> If you add something like:
> 
> 		r |= (CPU_UPAID << INTMAP_TID_SHIFT);
> 
> (after you figure out where those are defined) it will probably solve your 
> problem.

Perfect.  Thanks!  They are already included (ctlreg.h and iommureg.h).
Adding exactly that made my E3500 work with serial console, and with
keyboard/screen console.

I've put the kernel that I used to test with at:

  http://ftp.netbsd.org/pub/NetBSD/misc/jdc/netbsd-Exx00.gz

If compiling, update to revision 1.2 of src/sys/arch/sparc64/dev/fhc.c.

Thanks,

J

--

-- 
  My other computer also runs NetBSD    /        Sailing at Newbiggin

(Continue reading)

Michael Thompson | 30 Jul 2011 02:29

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Thu, Jul 28, 2011 at 10:27 PM, matthew green <mrg <at> eterna.com.au> wrote:
>
>> avail memory = 981 MB
>> mainbus0 (root): SUNW,Ultra-Enterprise (8-slot Sun Enterprise
>> E4500/E5500): hostid 809db71e
>> cpu0 at mainbus0: SUNW,UltraSPARC-II  <at>  336 MHz, UPA id 0
>> cpu0: 16K instruction (32 b/l), 16K data (32 b/l), 4096K external (64 b/l)
>> cpu at mainbus0 not configured
>> central at mainbus0 not configured
>> fhc at mainbus0 not configured
>
> i think one of these is the problem.  either central or fhc should
> be attaching something here, and we're missing some of the onboard
> devices attached here.  both serial ports and keyboard are missing
> from these boot logs.
>
> the reason it works with serial console is that pcons attaches.  the
> reason it doesn't work with the keyboard attached is that pcons does
> not attach here, presumably because wsdiplay0 claimed the console?
>
> can you send or link to a prtconf -v or whatever it is from solaris
> that will show the entire OFW device tree?  it would be great to get
> these systems finally supported!
>
> thanks.
>
>
> .mrg.
>

(Continue reading)

Eduardo Horvath | 1 Aug 2011 17:24

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Fri, 29 Jul 2011, Michael Thompson wrote:

> > can you send or link to a prtconf -v or whatever it is from solaris
> > that will show the entire OFW device tree?  it would be great to get
> > these systems finally supported!
> >
> > thanks.
> >
> >
> > .mrg.
> >
> 
> # prtconf
> System Configuration:  Sun Microsystems  sun4u
> Memory size: 11264 Megabytes
> System Peripherals (Software Nodes):

Actually, `prtconf -pv' (or is it `-pV'?)  would be better.

Edaurdo

Michael Thompson | 2 Aug 2011 01:55

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Mon, Aug 1, 2011 at 11:24 AM, Eduardo Horvath <eeh <at> netbsd.org> wrote:
> On Fri, 29 Jul 2011, Michael Thompson wrote:
>
>> > can you send or link to a prtconf -v or whatever it is from solaris
>> > that will show the entire OFW device tree?  it would be great to get
>> > these systems finally supported!
>> >
>> > thanks.
>> >
>> >
>> > .mrg.
>> >
>>
>> # prtconf
>> System Configuration:  Sun Microsystems  sun4u
>> Memory size: 11264 Megabytes
>> System Peripherals (Software Nodes):
>
> Actually, `prtconf -pv' (or is it `-pV'?)  would be better.
>
> Edaurdo

# prtconf -pV
OBP 3.2.30 2002/10/25 14:03

With just a single CPU to make the output a reasonable size...
# prtconf -pv
System Configuration:  Sun Microsystems  sun4u
Memory size: 4096 Megabytes
System Peripherals (PROM Nodes):

(Continue reading)

Michael Thompson | 30 Jul 2011 02:36

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Thu, Jul 28, 2011 at 10:27 PM, matthew green <mrg <at> eterna.com.au> wrote:
>
>> avail memory = 981 MB
>> mainbus0 (root): SUNW,Ultra-Enterprise (8-slot Sun Enterprise
>> E4500/E5500): hostid 809db71e
>> cpu0 at mainbus0: SUNW,UltraSPARC-II  <at>  336 MHz, UPA id 0
>> cpu0: 16K instruction (32 b/l), 16K data (32 b/l), 4096K external (64 b/l)
>> cpu at mainbus0 not configured
>> central at mainbus0 not configured
>> fhc at mainbus0 not configured
>
> i think one of these is the problem.  either central or fhc should
> be attaching something here, and we're missing some of the onboard
> devices attached here.  both serial ports and keyboard are missing
> from these boot logs.
>
> the reason it works with serial console is that pcons attaches.  the
> reason it doesn't work with the keyboard attached is that pcons does
> not attach here, presumably because wsdiplay0 claimed the console?
>
> can you send or link to a prtconf -v or whatever it is from solaris
> that will show the entire OFW device tree?  it would be great to get
> these systems finally supported!
>
> thanks.
>
>
> .mrg.
>

(Continue reading)

Michael Thompson | 30 Jul 2011 02:39

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Thu, Jul 28, 2011 at 10:27 PM, matthew green <mrg <at> eterna.com.au> wrote:
>
>> avail memory = 981 MB
>> mainbus0 (root): SUNW,Ultra-Enterprise (8-slot Sun Enterprise
>> E4500/E5500): hostid 809db71e
>> cpu0 at mainbus0: SUNW,UltraSPARC-II  <at>  336 MHz, UPA id 0
>> cpu0: 16K instruction (32 b/l), 16K data (32 b/l), 4096K external (64 b/l)
>> cpu at mainbus0 not configured
>> central at mainbus0 not configured
>> fhc at mainbus0 not configured
>
> i think one of these is the problem.  either central or fhc should
> be attaching something here, and we're missing some of the onboard
> devices attached here.  both serial ports and keyboard are missing
> from these boot logs.
>
> the reason it works with serial console is that pcons attaches.  the
> reason it doesn't work with the keyboard attached is that pcons does
> not attach here, presumably because wsdiplay0 claimed the console?
>
> can you send or link to a prtconf -v or whatever it is from solaris
> that will show the entire OFW device tree?  it would be great to get
> these systems finally supported!
>
> thanks.
>
>
> .mrg.
>

(Continue reading)

Eduardo Horvath | 28 Jul 2011 17:45

Picon

Re: Keyboard Problem with V5.1 on an E4000

On Wed, 27 Jul 2011, Michael Thompson wrote:

> I am trying to revive an E4000. I did a "boot cdrom" at the OpenBoot
> prompt, NetBSD boots, displays a bunch of configuration messages on
> the console, changes the configuration of the video display (it says
> wsdisplay0) displays the configuration messages, and asks for the
> terminal type. At this point the keyboard stops working.  A "stop-A"
> will not get it back to OpenBoot. They keyboard works OK with
> OpenSolaris.
> 
> Ideas?

I expect the standard kernel config doesn't support the driver attachment 
points you need.  

How is your keyboard attached to your E4K?  What's the OBP device path for 
the keyboard and display?  Do you have SBus or PCI I/O cards?

Eduardo

Return

Return to gmane.os.netbsd.ports.sparc64.

Search Archive

Language

Change language

Options

Current view: Threads only / Showing only 20 lines / Not hiding cited text.
Change to All messages, whole messages, or hide cited text.

Post a message
NNTP Newsgroup
Classic Gmane web interface
XML

XML

RSS Feed
List Information

About Gmane