Havard Eidnes | 1 Jan 2009 02:47
Picon

Package binaries for NetBSD/sparc64 3.1 / pkgsrc-2008Q3

Hi,

I've uploaded the results of a bulk build for NetBSD/sparc64 3.1
(compiled on 3.1_STABLE) to

   ftp://ftp.NetBSD.org/pub/pkgsrc/packages/NetBSD/sparc64/3.1_2008Q3/

This build was done from sources updated Nov 1 2008.  A total of
6.1GB of packages were uploaded, for a total of 6475 packages.

I will be switching this machine to the 2008Q4 branch now, so this
is the first and last update of 2008Q3 for this os/arch combination.

In keeping with established policy, which states to keep at most two
full sets of packages per arch/os/os-version, the 2008Q1 packages have
been removed for NetBSD/sparc64 version 3.1.

As indicated above, the build was done on a machine running
NetBSD 3.1_STABLE, and most of these packages should be usable both
on 3.0 and later 3.x versions of the OS.

Regards,

- Håvard

Havard Eidnes | 4 Jan 2009 12:07
Picon

Re: Package binaries for NetBSD/sparc64 4.0 / pkgsrc-2008Q3

Hi,

I've uploaded the updates from a bulk build for NetBSD/sparc64 4.0
(compiled on 4.0_BETA2 and subsequently 4.0.1) to

   ftp://ftp.NetBSD.org/pub/pkgsrc/packages/NetBSD/sparc64/4.0_2008Q3/

This build was done from sources updated Dec 4 2008.  A total of
3.4GB of packages were uploaded, for a total of 3080 packages.

Packages available in new versions or which were not available earlier
are:

aegis-4.22nb2, asterisk-1.2.30, btpd-0.12nb3, bugzilla-2.22.6,
bugzilla-3.0.6, clamav-0.94.2, compat14-1.4.3.1nb2, compat15-1.5.2,
compat16-1.6.2, compat20-2.1, compat30-3.1, courier-authlib-0.61.0,
crossfire-maps-1.9.0.1nb1, cups-1.3.9, dar-2.3.8, dovecot-1.1.6,
drupal-5.12, drupal-6.6nb1, ed-1.1, esniper-2.19.0, ettercap-0.6.bnb4,
fastjar-0.93nb3, firefox-2.0.0.18, firefox-gtk1-2.0.0.18,
freezetag-0.9.2nb4, gnetlist-1.4.0, graphviz-2.16.1nb3, horde-3.2.2,
imap-uw-2007dnb1, imap-uw-utils-20050108nb2, imlib2-1.4.2nb1,
imm-3.6a1nb1, imp-4.2nb1, imwheel-0.9.9nb1, ja-trac-0.11.1pl2,
jhead-2.84, koth-0.8.0nb4, ktorrent-2.2.8, libspf2-1.2.8,
libxml2-2.7.2nb2, lsof-4.78.4.0.1nb3, mailscanner-4.30.3.2nb3,
mailsync-5.2.1nb4, mantis-1.1.4, mp3asm-0.1.3.1, myodbc-3.51.26r1127,
mysql-client-4.1.22nb3, mysql-client-5.0.67nb1, mysql-server-4.1.22nb2,
nagios-base-3.0.3, ncftp3-3.1.9, net-snmp-5.4.2, netatalk-1.6.4anb5,
netatalk-2.0.3nb11, optipng-0.6.1, palmpower-1.0.2, pcemu-1.01anb3,
pcre++-0.9.5nb1, php-5.2.6nb2, php5-imap-5.2.6nb4, phpmyadmin-2.11.9.3,
pine-4.64nb5, png-1.2.33rc02, prayer-1.0.18nb5, py24-musique-0.5nb5,
(Continue reading)

Martin Husemann | 5 Jan 2009 11:05
Picon

Re: Firefox and creator problems

FYI: I fixed a nasty 64bit ABI bug in the call stubs (when C++ code invokes
methods of an object implemented in javascript) and now everything I tried
seems to work fine, including url bar, history and about:config view.

You need the latest + greatest firefox3 pkg for this, FF2 is no longer 
supported upstream, so I don't think it's worth to backport changes there.

Please let me know if you find something reproducable not working (that works
with other archs) or crashes. Keep the firefox-bin.core around - that (and
the binary in /usr/pkg/lib/firefox3) should be enough to debug the typical
alignement bus errors.

Martin

Jochen Kunz | 5 Jan 2009 13:39
Picon

Re: Firefox and creator problems

On Mon, 5 Jan 2009 11:05:13 +0100
Martin Husemann <martin <at> duskware.de> wrote:

> FYI: I fixed a nasty 64bit ABI bug in the call stubs
Great. I'll try later. At the moment I don't have time to rebuild my sparc64 machine.

> You need the latest + greatest firefox3 pkg for this,
Is this in pkgsrc-2008Q4?
If not is it planed to be pulled up to pkgsrc-2008Q4?
--

-- 

tschüß,
       Jochen

Homepage: http://www.unixag-kl.fh-kl.de/~jkunz/

Martin Husemann | 5 Jan 2009 13:49
Picon

Re: Firefox and creator problems

On Mon, Jan 05, 2009 at 01:39:32PM +0100, Jochen Kunz wrote:
> Is this in pkgsrc-2008Q4?

No.

> If not is it planed to be pulled up to pkgsrc-2008Q4?

Not sure yet if it's worth - we'll probably find more problems soon.

Martin

Matthias Scheler | 5 Jan 2009 13:52
Picon
Favicon

Re: Firefox and creator problems

On Mon, Jan 05, 2009 at 01:39:32PM +0100, Jochen Kunz wrote:
> > You need the latest + greatest firefox3 pkg for this,
> Is this in pkgsrc-2008Q4?

No, it isn't.

	Kind regards

--

-- 
Matthias Scheler                                  http://zhadum.org.uk/

Gert Doering | 7 Jan 2009 11:11
Picon

SATAlink PCI card and Ultra 5?

Hi,

(This is likely not to be NetBSD-related at all, but I hope that there is
enough hardware expertise with Sun hardware here to help figure out the
answer...)

I'm mystified... I'm trying to get SATA to my Ultra5, and for that purpose
got a number of "cheap off-the shelf" PCI-to-SATA controllers, to see which
one works with NetBSD/Sparc.

The one that looks most promising is SATAlink based, and it works perfectly
well (!!) in my Ultra60, with NetBSD-5.0:

satalink0 at pci0 dev 5 function 0
satalink0: Silicon Image SATALink 3512 (rev. 0x01)
satalink0: SATALink BA5 register space disabled
satalink0: bus-master DMA support present
satalink0: primary channel wired to native-PCI mode
satalink0: using ivec 1c for native-PCI interrupt
atabus0 at satalink0 channel 0
satalink0: secondary channel wired to native-PCI mode
atabus1 at satalink0 channel 1

(no devices attached right now, due to the Ultra60 not being suited very
well to power non-SCA drives - but I've tested it with a WD 750G drive, and
everything behaved quite normally).

Now the funny part.  The "target" machine for this controller is an Ultra5,
which direly needs a larger and faster hard disk.  PCI being PCI, I assumed
it would "just work".  But it doesn't.
(Continue reading)

Gert Doering | 7 Jan 2009 18:24
Picon

Re: SATAlink PCI card and Ultra 5?

Hi,

to followup with a bit more detail on this:

On Wed, Jan 07, 2009 at 11:11:26AM +0100, Gert Doering wrote:
> Now the funny part.  The "target" machine for this controller is an Ultra5,
> which direly needs a larger and faster hard disk.  PCI being PCI, I assumed
> it would "just work".  But it doesn't.
> 
>  - OBP does not recognize the card ("setenv diag-switch? true" shows the
>    PCI slot in question as "empty" - the slot is OK, though, as a PCI
>    RTL8019 network card is recognized just fine).

It's not actually a *Sun* hardware issue.  

I have now tested the card in a PC mainboard (Intel 440BX, P3-600, about 
the same age as the U5...) and the effect is quite similar - no Bios 
messages, and Linux' "lspci" does not see the card at all.

So it might be the card that is "funny" - it actually looks funny, because
it's a bit shorter than a normal PCI card - the last 4 rows of connection
pins are just "not there" (A59..A62 and B59..B62).

According to http://pinouts.ru/Slots/PCI_pinout.shtml these shouldn't
be needed in any way, though (additional voltage pins, plus 64bit 
signalling, which is likely to be not connected on the U5 anyway).

... moving back the card to the U60, just to be sure I haven't broken
anything, it Just Works... (just copied over a few gbytes to the SATA
disk, just to be sure).
(Continue reading)

Eduardo Horvath | 7 Jan 2009 18:25
Picon

Re: SATAlink PCI card and Ultra 5?


Whatcha need to do is figure out why OBP isn't seeing the device on the 
U5.  Turn on fcode-debug? and diag-switch?, reset the machine, cd to the 
/pci node and see if there's a "pci-debug" word visible.  If there is,
set it to true and type probe-all.  It should display what OBP does to
the BARs for each device it probes.

Alternatively, you can select the /pci node and config-l <at>  with the correct
config space address to directly read the device registers.  

If you can see the device then OBP is not probing it properly.  You should
be able to create an nvramrc script to generate a node for your device, or
possibly tweak the PCI prober to recognize the device.

If you can't see the config space registers then it's some sort of 
hardware incompatibility and you probably need to find a different card.

Eduardo

Gert Doering | 7 Jan 2009 18:45
Picon

Re: SATAlink PCI card and Ultra 5?

Hi,

On Wed, Jan 07, 2009 at 05:25:47PM +0000, Eduardo Horvath wrote:
> Whatcha need to do is figure out why OBP isn't seeing the device on the 
> U5.  Turn on fcode-debug? and diag-switch?, reset the machine, cd to the 
> /pci node and see if there's a "pci-debug" word visible.  If there is,
> set it to true and type probe-all.  It should display what OBP does to
> the BARs for each device it probes.

Ummm.  You've lost me here.  I know how to set "diag-switch?", but I
have no idea how to check for the "pci-debug" word, and how to set it...

> Alternatively, you can select the /pci node and config-l <at>  with the correct
> config space address to directly read the device registers.  

... ditto... :(

Could you show me sample command lines, and the to-be-expected output?

(Alternatively: is there a good online documentation about this OBP stuff
available?)

> If you can see the device then OBP is not probing it properly.  You should
> be able to create an nvramrc script to generate a node for your device, or
> possibly tweak the PCI prober to recognize the device.
> 
> If you can't see the config space registers then it's some sort of 
> hardware incompatibility and you probably need to find a different card.

Given the fact that it doesn't work in the Pentium-III PC either, it might
(Continue reading)


Gmane