Re: [osol-discuss] Idea for a nice OpenSolaris project - HCL client

2006-01-17 Thread Joerg Schilling
Erast Benson [EMAIL PROTECTED] wrote:

 On Mon, 2006-01-16 at 14:38 +0100, Joerg Schilling wrote:
  Bruce Riddle [EMAIL PROTECTED] wrote:
  
   How about a really intelligent parser of prtconf -pv output.
  
  Dan Micks prtpci does already a lot

 I know about this wrapper. But lspci is what every GNU/Linux system has
 today and it is fully ported to Solaris, which makes prtpci obsolete.
 IMHO

I cannot find a port for Solaris.

And BTW: I would guess that lspci (in contrary to prtpci) needs root 
privileges..

Jörg

-- 
 EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
   [EMAIL PROTECTED](uni)  
   [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] re: [networking-discuss] Re: Re: The 802.3ad link aggregationsupport onnetra 240 (bge broadcom) under solaris 8

2006-01-17 Thread Peter Memishian

  but I am not very clear the meaning of   By moving it into the networking
  stack,  can I consider the way realize the 802.3ad features in sol 10 is
  just moving it from kernel to networking stack?

The networking stack is in the kernel.

With Sun Trunking, the feature is part of the drivers -- e.g., coded into
the ce, qfe, and ge drivers (along with a related 'laggr' module).  With
the bundled support, the feature is part of the GLDv3 kernel networking
framework.  This means that any driver written to GLDv3 will support link
aggregation.  Further, as part of Clearview, we are developing a shim that
will allow any driver to appear as a GLDv3 driver -- and thus allow link
aggregation to work with any Ethernet driver.

--
meem
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] re: [networking-discuss] The 802.3ad link aggregation support on netra 240 (bge broadcom) under solaris 8

2006-01-17 Thread Peter Memishian

  After checking some blog pages of open solaris, I knew that the old type
  drivers of NIC adapter (eg. qe,fqe,ge) was developed by DLPI or GLDv2, but
  the bge was developed by GLDv3. I guess sun trunking only support the former
  method.

Sun Trunking is supported by some extensions coded directly into some of
the existing monolithic DLPI drivers, such as ge or qfe.  The approach has
significant flexibility issues (e.g., you can't trunk across different
drives), and thus we have moved away from it.

  Sun does provide a solution: migrate the platform to solaris 10 and use the
  built-in command dladm to realize the 802.3 link aggregation support. It's
  no sense to me, there are too many 3-party applications developed under
  solaris 8 in this project, it's impossible to migrate all of them to solaris
  10 just because of the link aggregation problem.

If you're stuck on Solaris 8, your options are limited.  There are many
reasons besides built-in link aggregation to move to Solaris 10, though.
We also take our binary compatibility guarantee seriously, and would also
be interested in knowing what problems you encounter with third-party
software.  If it is a matter of vendor support, we would also like to know
what key products you need are not supported on Solaris 10.

  Any one knows which 3-party software can provide this kind of function:

I'm not aware of one.  However, I'd think you could get some Sun GigaSwift
Ethernet (ce) cards, and use Sun Trunking to aggregate them.

  realize 802.3 support under solaris 8 by GLDv3 method, otherwise I have to
  study the solaris 10 code for link aggregation and port them to sol 8.

I think that would be extremely difficult -- and you would of course end
up with a completely unsupported solution.

-- 
meem
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] re: [networking-discuss] The 802.3ad link aggregation support on netra 240 (bge broadcom) under solaris 8

2006-01-17 Thread Peter Memishian

  Sun Trunking is supported by some extensions coded directly into some of
  the existing monolithic DLPI drivers, such as ge or qfe.  The approach has
  significant flexibility issues (e.g., you can't trunk across different
  drives), and thus we have moved away from it.

That should read different drivers -- e.g., you can't create a trunk
containing ce0 and ge0.

--
meem
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] re: [networking-discuss] Re: The 802.3ad link aggregation support onnetra 240 (bge broadcom) under solaris 8

2006-01-17 Thread Peter Memishian

  Thanks for you guys' immediate answer, just like you mentioned, up to now
  it's an availabe way to plug the ce cards in the netra 240 boxes and use the
  sun trunking. 

... and this is no longer an option?  AFAIK, ce cards are still available
and qualified for a Netra 240.   It's worth pointing out that this issue
is because a core OS feature (802.3ad support) was built into the driver
rather than the networking stack.  By moving it into the networking stack,
we ensure this will not happen again.

--
meem
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] Idea for a nice OpenSolaris project - HCL client

2006-01-17 Thread Erast Benson
On Tue, 2006-01-17 at 11:50 +0100, Joerg Schilling wrote:
 Erast Benson [EMAIL PROTECTED] wrote:
 
  On Mon, 2006-01-16 at 14:38 +0100, Joerg Schilling wrote:
   Bruce Riddle [EMAIL PROTECTED] wrote:
   
How about a really intelligent parser of prtconf -pv output.
   
   Dan Micks prtpci does already a lot
 
  I know about this wrapper. But lspci is what every GNU/Linux system has
  today and it is fully ported to Solaris, which makes prtpci obsolete.
  IMHO
 
 I cannot find a port for Solaris.

you could grab it from Nexenta SVN.

 And BTW: I would guess that lspci (in contrary to prtpci) needs root 
 privileges..

True. But main point is popularity:

Google:

lspci - 611,000
prtconf - 79,100
prtpci - 201

___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] KDE 3.4.3 Source Code at Sunfreeware

2006-01-17 Thread Alan DuBoff
On Saturday 14 January 2006 10:21 pm, Stefan Teleman wrote:
 The entire source code tree for the KDE 3.4.3 port is now available
 for download from Sunfreeware and its mirrors:

 http://www.sunfreeware.com/kde.html

 Many Thanks to Steve Christensen for taking the time to set this up
 and for making the full source code available.

 Final source code patches for K3B will follow very soon.

Thanks to both you and SteveC for working that out together!

The community will appreciate it!

-- 

Alan DuBoff - Sun Microsystems
Solaris x86 Engineering


___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] ACPI PCI-E issues installing on HP NX9600 laptop

2006-01-17 Thread Alan DuBoff
On Sunday 15 January 2006 02:10 pm, Bruce Riddle wrote:
 I'm having no luck installing Solaris express on my HP nx9600.

This is something that will need to be looked at by either Jan or Dana. Since 
we've exhausted most of what I can think of, I don't know what's wrong with 
it. Hopefully one of them can help you.

-- 

Alan DuBoff - Sun Microsystems
Solaris x86 Engineering


___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] KDE 3.4.3 Source Code at Sunfreeware

2006-01-17 Thread Alan DuBoff
On Monday 16 January 2006 11:58 am, Erast Benson wrote:
 I'd like to apply your changes to Nexenta KDE.

Erast,

That would be awesome!

-- 

Alan DuBoff - Sun Microsystems
Solaris x86 Engineering


___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] New Community Proposal: Games

2006-01-17 Thread Bart Smaalders

ken mays wrote:

Utilize the Nvidia Developer, NVZone, and Solaris
forums while you are at it

Crashing up GLTron lightcycles is a favorite pastime
of mine (http://www.gltron.org/)...

~ Ken Mays


Having worked on the original Doom and Quake ports to Solaris,
I think this is a great idea :-)!  I'd really like to get more
games available for Solaris, but I haven't had time to play with
this much lately.

I did manage to get the Really Slick Screensavers working
on Solaris x86... works pretty well if you have a good graphics
card. http://rss-glx.sourceforge.net/

- Bart

--
Bart Smaalders  Solaris Kernel Performance
[EMAIL PROTECTED]   http://blogs.sun.com/barts
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] New Community Proposal: Games

2006-01-17 Thread Rich Teer
On Tue, 17 Jan 2006, Bart Smaalders wrote:

 Having worked on the original Doom and Quake ports to Solaris,
 I think this is a great idea :-)!  I'd really like to get more

+1

Which reminds me: any chance of getting Doom Arena working on
newer versions of Solaris?  ISTR having problems last time I
tried (S8 or S9 on an Ultra 1 with Creator 3D frame buffer).

-- 
Rich Teer, SCNA, SCSA, OpenSolaris CAB member

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-group.com/rich
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] New Community Proposal: Games

2006-01-17 Thread Bart Smaalders

Rich Teer wrote:

On Tue, 17 Jan 2006, Bart Smaalders wrote:


Having worked on the original Doom and Quake ports to Solaris,
I think this is a great idea :-)!  I'd really like to get more


+1

Which reminds me: any chance of getting Doom Arena working on
newer versions of Solaris?  ISTR having problems last time I
tried (S8 or S9 on an Ultra 1 with Creator 3D frame buffer).



Wow... I didn't know anyone was still running that.  Newer
machines can run doom fine w/o needing sundgadoom; that version
was always difficult because it has coded into it the drivers
for the various framebuffers available at the time (on sparc).

When I whacked on quake I used libxil, which took care of
the direct graphics access and the pixel multiple in a much
nicer way, albeit with a bit more overhead.

The authors of Doomarea were Doug Stein and Steve Jankowski;
it's a tcl script if I remember correctly.  I know less than
nothing about tcl, but it should still work.  You will need
to replace the sundgdoom executable w/ prboom or similar from
blastwave.

- Bart



--
Bart Smaalders  Solaris Kernel Performance
[EMAIL PROTECTED]   http://blogs.sun.com/barts
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] Reverting to old build: Solaris Express - Community Release

2006-01-17 Thread Karyn Ritter
I was just notified of a new legal issue with build 30 of the Solaris 
Express - Community Release. We hope that this is a temporary problem, 
and that we'll be able to re-post the images for this build shortly.


In the meantime, I've changed the links on the download page back to 
build 28.


Because we're posting these more quickly than we've ever released 
Solaris Express builds, we're bound to encounter some hiccups. I'm 
hoping that there will be fewer of these in the future. I apologize for 
the inconvenience, and appreciate your understanding.


Thanks,

Karyn
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] CIFS group creation request.

2006-01-17 Thread Stephen Hahn

   Speaking in the most bureaucratic manner I can muster, you need to
   decide whether your efforts require a project or a community.  If
   your intent is to produce code or a structured best practice
   document or some other form of collaboratively authored product, then
   I recommend that you request a CIFS project, in which case you need
   to send mail to this alias like that of the network-sip project 

   http://www.opensolaris.org/jive/thread.jspa?threadID=4968tstart=0

   and find a seconder.
   
   If you believe that many CIFS projects might spawn from a long and
   wide-ranging discussion, and that such a discussion should be
   distinct from those of the SVM, UFS, and ZFS communities, then you
   need to articulate by email to this alias how a CIFS community is
   needed, meaningful, and separate, similar to the recent Appliances
   community proposal

   http://www.opensolaris.org/jive/thread.jspa?threadID=4945tstart=0

   (which, I note, mentions a zfs-Samba combination in the thread) and
   then either consensus, including CAB assent, or criticism will
   emerge.

   - Stephen
   
-- 
Stephen Hahn, PhD  Solaris Kernel Development, Sun Microsystems
[EMAIL PROTECTED]  http://blogs.sun.com/sch/
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] Re: Reverting to old build: Solaris Express - CommunityRelease

2006-01-17 Thread Peter Lees
is there no chance of a build29 release ?
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] Reverting to old build: Solaris Express - Community Release

2006-01-17 Thread Dennis Clarke
On 1/17/06, Karyn Ritter [EMAIL PROTECTED] wrote:
 I was just notified of a new legal issue with build 30 of the Solaris
 Express - Community Release. We hope that this is a temporary problem,
 and that we'll be able to re-post the images for this build shortly.


The source file opensolaris-src-20060102.tar.bz2  is OKay ?

Please please say yes.

Dennis
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Re: [osol-discuss] Build times for Open Solaris....

2006-01-17 Thread Dennis Clarke
On 1/17/06, John Kaitschuck [EMAIL PROTECTED] wrote:


 I have noticed that full build times seem to be a bit
 on the excessive side for Open Solaris on my older
 U60. Excessive here means on the order of more than
 8 hours. I was wondering what are the full build
 times for something like the AMD SMP (2 way) based
 Sun 2100z? or another more modern 64 bit architecture?

 I want to get this down to something a little more
 reasonable, vs. what currently equates to an over
 night build. While I have some money to throw at
 this problem, in the guise of new hardware, I am
 not in a position to buy a 4 way, SCSI based system.

 Does any one have any recommendations and real world
 build times to report?


Dual AMD 250 Opteron = about 1 hour build times
so .. fire off nightly .. watch a Star Trek re-run.
add 10 minutes to that and you have rebooted and are running fine.

Dennis
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Re: [osol-discuss] Build times for Open Solaris....

2006-01-17 Thread Joerg Schilling
Dennis Clarke [EMAIL PROTECTED] wrote:

 Dual AMD 250 Opteron = about 1 hour build times
 so .. fire off nightly .. watch a Star Trek re-run.
 add 10 minutes to that and you have rebooted and are running fine.

What machine is this?

Interesting: dual AMD 248 V20z  1:30 
 dual AMD 242 MSI-9617 (1GB RAM)1:30

Jörg

-- 
 EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
   [EMAIL PROTECTED](uni)  
   [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] Build times for Open Solaris....

2006-01-17 Thread Dennis Clarke
On 1/17/06, Joerg Schilling [EMAIL PROTECTED] wrote:
 Dennis Clarke [EMAIL PROTECTED] wrote:

  Dual AMD 250 Opteron = about 1 hour build times
  so .. fire off nightly .. watch a Star Trek re-run.
  add 10 minutes to that and you have rebooted and are running fine.

 What machine is this?

 Interesting: dual AMD 248 V20z  1:30
  dual AMD 242 MSI-9617 (1GB RAM)1:30


$ cat log/log.1223/mail_msg

 Nightly distributed build started:   Thu Dec 22 23:46:04 EST 2005 
 Nightly distributed build completed: Fri Dec 23 01:05:33 EST 2005 

 Total build time 

real1:19:29

that was build 29 and build 30 was about the same.  Build 28 was no
joy and build 27a was real fast at about 1 hour flat.

Dual AMD 250  V20z

Dennis
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Re: [osol-discuss] Build times for Open Solaris....

2006-01-17 Thread Bill Rushmore
On Tue, 2006-01-17 at 19:05, John Kaitschuck wrote:

 Does any one have any recommendations and real world
 build times to report?
 

Sun Blade 100: 8-10 Hours
SUNPCI III As guest Vmware OS on Windows XP off of a USB drive:  11+
hours
Sony Vaio TR1:  6 hours

Remember a U60 is an old system and that build time doesn't seem to bad
for its age.  
My recommendation are to use the -i option for an incremental build
whenever you can.  If you are working a specific section of the code
just build that portion while you are developing and save the complete
build for when you are done for the day.  Heck that is probably why it's
called 'nightly'. :-)

Bill
rushmores.net
 

___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] Build times for Open Solaris....

2006-01-17 Thread Bart Smaalders

Bill Rushmore wrote:

On Tue, 2006-01-17 at 19:05, John Kaitschuck wrote:


Does any one have any recommendations and real world
build times to report?



Sun Blade 100: 8-10 Hours
SUNPCI III As guest Vmware OS on Windows XP off of a USB drive:  11+
hours
Sony Vaio TR1:  6 hours

Remember a U60 is an old system and that build time doesn't seem to bad
for its age.  
My recommendation are to use the -i option for an incremental build

whenever you can.  If you are working a specific section of the code
just build that portion while you are developing and save the complete
build for when you are done for the day.  Heck that is probably why it's
called 'nightly'. :-)

Bill
rushmores.net
 


Note that dmake is a big win, even for single cpu machines since
it allows you to do something useful while waiting for the disk.

SPARC builds involve several separate kernel architectures, x86 just has
32 and 64 bit.

There are significant parts of the build that are rather serial;
Amdahl's law is alive and well in our builds.

- Bart



--
Bart Smaalders  Solaris Kernel Performance
[EMAIL PROTECTED]   http://blogs.sun.com/barts
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


Re: [osol-discuss] Build times for Open Solaris....

2006-01-17 Thread Matt Ingenthron

Bart Smaalders wrote:


There are significant parts of the build that are rather serial;
Amdahl's law is alive and well in our builds.


Careful there...  Since the guy that threw a wrench into Amdahl's law in 
1988 does research at Sun.  :)  Though I did turn up something in a 
search just now that says the two are mathematically equivalent.


Though, if I understand Gustafson's law correctly, you'll need to be 
building more kernels to be able to use that style parallelism.  It's 
more appropriate for HPC algorithms, not building kernels.  :)


- Matt
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org


[osol-discuss] Found a gr8 website

2006-01-17 Thread Vicky Becker
Found a site several days ago and by using their affiliate program I can buy 
tickets to Las Vegas shows for less than the cost at the casino box office.  
www.insidervlv.com
This message posted from opensolaris.org
___
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org