headers
Chris Ross | 9 Feb 2010 20:07

dbri and speakerbox on a SPARCstation 10


   I just updated an old SPARCstation 10 build machine (3 processors)  
to netbsd-5.  Upon booting, I happened to notice:

dbri0: speakerbox detected
dbri0: cs4215 rev E found at offset 8
audio0 at dbri0: full duplex, playback, capture, mmap

   Now, while it's cool that the audio is supported, even if I don't  
expect I'll use it, I find the first line odd.  I'm sure there isn't a  
speakerbox on this machine.  I don't even remember which incarnations  
of SPARC's used the little speakerboxes (18-20 years ago), so maybe  
the 10 has an effective speakerbox in it that's being detected, but I  
think the description, if not the fact, is somewhat wrong.

   Just curious, anyone know why that appears?

                   - Chris
Permalink | Reply | 2 comments |
headers
Michael | 9 Feb 2010 22:00
Picon

Re: dbri and speakerbox on a SPARCstation 10


Hello,

On Feb 9, 2010, at 2:07 PM, Chris Ross wrote:

>  I just updated an old SPARCstation 10 build machine (3 processors)  
> to netbsd-5.  Upon booting, I happened to notice:
>
> dbri0: speakerbox detected
> dbri0: cs4215 rev E found at offset 8
> audio0 at dbri0: full duplex, playback, capture, mmap
>
>  Now, while it's cool that the audio is supported, even if I don't  
> expect I'll use it, I find the first line odd.  I'm sure there isn't  
> a speakerbox on this machine.  I don't even remember which  
> incarnations of SPARC's used the little speakerboxes (18-20 years  
> ago), so maybe the 10 has an effective speakerbox in it that's being  
> detected, but I think the description, if not the fact, is somewhat  
> wrong.
>
>  Just curious, anyone know why that appears?

The DAC sits on a synchronous serial interface from the ISDN chip,  
according to the docs I have it's on different time slots when onboard  
or in a speakerbox so we check both. Supposedly offset 8 is for  
external and 0 for onboard - I think I'll just remove that printf.

have fun
Michael
(Continue reading)

Permalink | Reply | 0 comments |
headers
John Ruschmeyer | 10 Feb 2010 18:15
Picon

Re: dbri and speakerbox on a SPARCstation 10

On Tue, 2010-02-09 at 14:07 -0500, Chris Ross wrote:
> I just updated an old SPARCstation 10 build machine (3 processors)  
> to netbsd-5.  Upon booting, I happened to notice:
> 
> dbri0: speakerbox detected
> dbri0: cs4215 rev E found at offset 8
> audio0 at dbri0: full duplex, playback, capture, mmap
> 
>    Now, while it's cool that the audio is supported, even if I don't  
> expect I'll use it, I find the first line odd.  I'm sure there isn't a  
> speakerbox on this machine.  I don't even remember which incarnations  
> of SPARC's used the little speakerboxes (18-20 years ago), so maybe  
> the 10 has an effective speakerbox in it that's being detected, but I  
> think the description, if not the fact, is somewhat wrong.

The speakerbox was an external device, so you'd probably know if you had
one. I know they were supported on the SPARCstation 10 and LX (I put one
on my LX after I removed the speaker to improve cooling. I don't think
they were supported on too many other systems, though.

John
Permalink | Reply | 0 comments |
headers
Chris Ross | 11 Feb 2010 20:34

Threaded lang/perl5 on sparc?


   Hello.  So, I'm updating a sparc image to 5.0_STABLE, and was  
rebuilding packages that I had on the previous image.  This includes  
perl (previously 5.8.8, now to be 5.10.1).

   The default (HEAD pkgsrc) perl configuration builds with threads,  
and I know there have been threading problems on sparc/sparc64 for a  
while now.  So, does anyone know whether I can presume threads might  
work alright in perl?  The prime reason I ask, is because after perl  
was built, a make test showed:

ext/threads-shared/t/0nothread................................ok
ext/threads-shared/t/av_refs..................................ok
ext/threads-shared/t/av_simple................................ok
ext/threads-shared/t/blessed..................................ok
ext/threads-shared/t/clone....................................ok
ext/threads-shared/t/cond.....................................ok
ext/threads-shared/t/disabled.................................ok
ext/threads-shared/t/hv_refs..................................ok
ext/threads-shared/t/hv_simple................................ok
ext/threads-shared/t/no_share.................................ok
ext/threads-shared/t/object...................................ok
ext/threads-shared/t/shared_attr..............................ok
ext/threads-shared/t/stress...................................# Test  
failed: 29 threads timed out
FAILED at test 1
ext/threads-shared/t/sv_refs..................................ok
ext/threads-shared/t/sv_simple................................ok
ext/threads-shared/t/utf8.....................................ok
ext/threads-shared/t/waithires................................ok
(Continue reading)

Permalink | Reply | 4 comments |
headers
Martin Husemann | 11 Feb 2010 20:42
Picon

Re: Threaded lang/perl5 on sparc?

On Thu, Feb 11, 2010 at 02:34:31PM -0500, Chris Ross wrote:
>   The default (HEAD pkgsrc) perl configuration builds with threads,  
> and I know there have been threading problems on sparc/sparc64 for a  
> while now.

There are no known problems after we changed to a 1:1 threading model
(before 5.0).

I'll look at the failing perl test.

Martin
Permalink | Reply | 3 comments |
headers
der Mouse | 14 Feb 2010 05:06

Re: aak! Tadpole 3GX parts/manual?

Back on 2010-01-25, I wrote

> I have a Tadpole SPARCbook 3GX.  Just today, I went to close the lid
> and, with an unpleasant cracking sound, the right hinge broke.  [...]

A kind listmember from another list sent me a replacement bezel (the
part of the hinge that broke is integral with the bezel), and I just
replaced it; the machine is now back in working order.

The replacement bezel came with hinges attached, and, now that I have
something to compare with, I realize the old hinges were very stiff;
this is probably how come the part that broke got overstressed to begin
with.  The new hinges move much more smoothly and with far less effort.

I've also taught myself to take the screen apart enough to replace the
bezel, of course; in the process, I've learnt that there's a connector
between a pc board and the LCD unit proper that, if slightly loose, can
cause a very odd failure mode: colours are weird and the display gets
shifted left roughtly 40 pixels and up maybe 10-15 pixels.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse <at> rodents-montreal.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B
Permalink | Reply | 0 comments |
headers
BERTRAND Joel | 14 Feb 2010 15:07
Picon

Re: [5.99.24] Async registers

	Hello,

BERTRAND Joel a écrit :
> New kernel bug : under high load, MP system returns to PROM prompt. I've
> never seen it with same kernel and only one CPU. Information printed on
> console :
>
> Async register (mid 8): afsr=0x0 <AFA=0x0> afva=0x00
> Watchdog reset
>
> System : dual RT626, but I'm not sure that this bug is HS specific.

	Any news ? I have tried to trigger this bug without any success. I 
don't understand why a hard condition is send to NMI without memory 
corruption. Watchdog reset occurs 4 or 5 seconds _after_ 'async 
register' message.

	Regards,

	JKB
Permalink | Reply | 0 comments |
headers
BERTRAND Joel | 14 Feb 2010 15:09
Picon

Re: Threaded lang/perl5 on sparc?

Martin Husemann a écrit :
> On Thu, Feb 11, 2010 at 02:34:31PM -0500, Chris Ross wrote:
>>    The default (HEAD pkgsrc) perl configuration builds with threads,
>> and I know there have been threading problems on sparc/sparc64 for a
>> while now.
>
> There are no known problems after we changed to a 1:1 threading model
> (before 5.0).
>
> I'll look at the failing perl test.

	On my SS20 MP and with 5.99.24 kernel/userland, failing test fails with 
well known kern_lwp error (cannot allocate).

	Regards,

	JKB
Permalink | Reply | 0 comments |
headers
Chris Ross | 16 Feb 2010 02:55

Re: Threaded lang/perl5 on sparc?


On Feb 11, 2010, at 14:42, Martin Husemann wrote:
> On Thu, Feb 11, 2010 at 02:34:31PM -0500, Chris Ross wrote:
>>  The default (HEAD pkgsrc) perl configuration builds with threads,
>> and I know there have been threading problems on sparc/sparc64 for a
>> while now.
>
> There are no known problems after we changed to a 1:1 threading model
> (before 5.0).
>
> I'll look at the failing perl test.

   Any update on this?  Have you had a chance to look into why this  
perl test case fails on a netbsd-5 sparc?

   Thanks.

            - Chris
Permalink | Reply | 1 comment |
headers
Alexander Breibach | 16 Feb 2010 11:19

post install crash after "dbri1: no mmcodec found" with NetBSD 5.0.2 on SparcStation20

Hello,

The error already reported port-sparc/41055 still exists in 5.0.2.
(http://mail-index.netbsd.org/netbsd-bugs/2009/03/22/msg009997.html)

I installed 5.0.2 -RELEASE/sparc on a SparcStation20.
After rebooting, the system stops after "dbri1: no mmcodec found".
This problem can be repeated by rebooting.

Any suggestions how to handle the problem?

If you need more information, please let me know.

Thanks in advance,
Alex
Permalink | Reply | 3 comments |

Gmane