headers
Paul | 1 Jan 2003 06:27
Picon
Picon
Favicon

can someone explain this pkgsrc thing please?

Hi,

In www/opera makefile it says:

ONLY_FOR_PLATFORM=	NetBSD-*-i386 SunOS-*-sparc

If I have sunos emulation does this mean I can 
make opera on sparc?
Or is it only for SunOS on sparc?

I'm quite confused about this. 
My make doesn't work, but it may not be meant to?

Thankyou.
Paul.

--

-- 
____________________________________________
Paul 
____________________________________________
Permalink | Reply | 1 comment |
headers
mcmahill | 1 Jan 2003 06:41
Picon
Favicon

Re: can someone explain this pkgsrc thing please?


It means that the package has not been set up for NetBSD-*-sparc even with
sunos emulation yet.  It may be that opera uses some bits of solaris
system calls which haven't been implemented in the emulation code or
simply that no one has tested out the pkg and enabled it for
NetBSD-*-sparc yet.

-Dan

On Wed, 1 Jan 2003, Paul wrote:

> Hi,
> 
> In www/opera makefile it says:
> 
> ONLY_FOR_PLATFORM=	NetBSD-*-i386 SunOS-*-sparc
> 
> If I have sunos emulation does this mean I can 
> make opera on sparc?
> Or is it only for SunOS on sparc?
> 
> 
> I'm quite confused about this. 
> My make doesn't work, but it may not be meant to?
> 
> 
> Thankyou.
> Paul.
> 
> --
(Continue reading)

Permalink | Reply | 0 comments |
headers
Martin Husemann | 3 Jan 2003 18:31
Picon

nell-users: heads up, config file change needed

If you are using the nell pcmcia bridge, please update your kernel config
file (note: this is for sparc nell users only, the sparc64 equivalent will
happen shortly, but SUN4U users are not affected yet).

The WI_AT_BIGENDIAN_HACK option is now obsolete, please remove it.
Instead the nell driver will insist on the presence of the new

options FULL_SPARC_BUS_SPACE

This makes the bus_space implementation complete, thanks to Mathew Green.
At this stage now nell should work with all pcmcia cards support by any
properly written MI netbsd driver. If you encounter any that does not
work, please send-pr.

In the few benchmarks Matthew and I did there was no real measurable 
performance hit by enabling FULL_SPARC_BUS_SPACE, but kernel size increases
by more than 30 kByte (haven't tested the difference on a full blown GENERIC
kernel yet).

That's why it's optional and disabled by default.

Martin
Permalink | Reply | 3 comments |
headers
John Steele Scott | 5 Jan 2003 12:04

a couple of low level questions


Hi all,

I've been playing around with sparc assembly on an IPC with NetBSD installed. 
I've learnt a lot over the last couple of days, but I have a couple of 
questions that someone knowledgeable could answer if they have time.

First, page 195 of the "SPARC Architecture Manual" seems to imply that when 
saving registers, it would be faster to go two registers at a time, using the 
std instruction instead of st. Yet in the code which NetBSD uses to save 
regs, st is used. Why is this?
<http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/arch/sparc/gen/saveregs.S?rev=1.3&content-type=text/x-cvsweb-markup>

My other question concerns the extra instructions in SPARCv8 which were not in 
SPARCv7, i.e. multiplication and division. The "SPARC Architecture Manual" 
proposes that code should call the ".udiv" routine (for example) which would 
then know whether to do the operation in software or hardware. Does this 
happen on NetBSD? Whereabouts in the libc source can I see this?

Thanks to anyone who takes the time to answer.

cheers,

John
Permalink | Reply | 0 comments |
headers
matthew green | 5 Jan 2003 12:51
Picon
Favicon

SMP success


hi folks.

due largely to the efforts of paul kranenburg, SMP support on sparc has
finally arrived.  this has only been tested successfully so far on an
ss20/712, an sun4/690 and an ss10/712 (i think), with a failure on a
4/670.  hypersparc support is almost certainly busted...

there are still some bugs and there are lots of implementation details
that could be reworked, but it seems pretty stable and performant so far.

see my boot log below.

.mrg.

<#0> ok boot netbsd.mp.234
Resetting ...
screen not found.
Can't open input device.
Keyboard not present.  Using tty for input and output.

SPARCstation 20 MP (2 X SuperSPARC-II), No Keyboard
ROM Rev. 2.25, 160 MB memory installed, Serial #3445063.
Ethernet address 8:0:20:20:23:48, Host ID: 72349147.

Rebooting with command: netbsd.mp.234
Boot device: /iommu/sbus/espdma <at> f,400000/esp <at> f,800000/sd <at> 3,0  File and args: netbsd.mp.234
>> NetBSD/sparc Secondary Boot, Revision 1.9
>> (mrg <at> russian-intervention.eterna.com.au, Mon Aug  6 01:11:16 EST 2001)
Booting netbsd.mp.234
(Continue reading)

Permalink | Reply | 18 comments |
headers
Walter Justen | 5 Jan 2003 13:25
Picon

Re: SMP success

Great news Matthew,

On Sun, 05 Jan 2003 22:51:12 +1100, matthew green wrote
>  hypersparc support is almost certainly busted...
>

do you have access to a hypersparc smp system for testing ? i can give you
console remote access via ssh to a 2x 180 MHz SS20 in the next days if needed ...

Walt
Permalink | Reply | 1 comment |
headers
matthew green | 5 Jan 2003 15:19
Picon
Favicon

re: SMP success


   On Sun, 05 Jan 2003 22:51:12 +1100, matthew green wrote
   >  hypersparc support is almost certainly busted...
   >

   do you have access to a hypersparc smp system for testing ? i can give you
   console remote access via ssh to a 2x 180 MHz SS20 in the next days if needed ...

someone with a MP hypersparc will be looking into this support...but thanks!
Permalink | Reply | 0 comments |
headers
Draco Paladine | 5 Jan 2003 16:25

How can one get a hold of hardware specifications for programming the SparcStation IPC?

Good day all:

	I have been playing around with the NetBSD port to the SparcStation
IPC for the past few days. I havent gotten around to installing the source
yet, but wondered:

1) how did the author of the port get a hold of the hardware specifications
   for the SparcStation IPC (as an example). There is a lot of information
   missing from Sun's manuals (especially that needed to write or port
   over an OS)

   a) memory map
   b) i/o ports and addressing
   c) sbus specification
   d) firmware info
   e) firmware structure in terms of what it does at power up and what
      and where it expects to find the bootstrap code
   f) structure of boot strap code
   g) register level/info/specifications for the framebuffer and network
      devices (sbus)

2) are these available from Sun?

Hopefully someone can shed some light on this -- thanks
--

-- 
In the names of Osiris, Isis and Seth
Let the Spirit of Anubis, Flow into THIS Vessel
Permalink | Reply | 1 comment |
headers
Erik E. Fair | 5 Jan 2003 22:44

Re: How can one get a hold of hardware specifications for programming the SparcStation IPC?

The Sun firmware is an early version of what became IEEE 1275 
(OpenBoot Firmware) which based on the Forth programming language. 
The idea is really quite clever: put a Forth interpreter on the 
motherboard, and add-in boards only need to have a driver written in 
Forth in their little PROMs, regarless of the system's processor - 
the same board with the same driver should work fine in a SPARC, an 
UltraSPARC, and, because Apple adopted IEEE 1275 for its PCI based 
PowerPC systems, in those too. It's a real shame that Intel/AMD/PC 
clone vendors didn't adopt it too.

See
	http://playground.sun.com/1275/

for some useful online docs.

The best documentation for Sun's Frame Buffers is in the Sun Frame Buffer FAQ:

	http://www.sunhelp.org/faq/FrameBuffer.html

Alas, this does not cover programming as such.

The workstations generally used AMD Local Area Network Controller - 
Ethernet (the famous LANCE) 7990, and docs for that are available 
from AMD's web site. See the le(4) man page for the last URL we had 
for that.

The serial ports are common Zilog 8530 DUARTs. This covers the mouse 
& keyboard too.

The SCSI is NCR (now Symbios) 5380.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Erik E. Fair | 5 Jan 2003 22:46

Re: SMP success

Congratulations! I've had a dual processor SS20 with 85 MHz 
SuperSPARC II's in it waiting for this day...

	but still looking forward to HyperSPARC,

	Erik <fair <at> clock.org>
Permalink | Reply | 0 comments |

Gmane