date:20031002


Antonello wrote:


On Thursday 02 October 2003 19:40, Ben Collins wrote:

 


/dev/scsi/host0/bus0/target1/lun0/part4 on /home type ext3 (rw)
 


That's a rather ugly partition map :)
Did you read the install docs where it says to not put a filesystem on
partition number 3 and leave it as a Whole Disk to make OBP happy?
   



RTFineM?!! Pfft!   :-)

Now that you've inspired me to at least glance at the documentation, I 
find only two references:


Sun disk partitions allow for 8 separate partitions (or slices). The 
third partition is usually (and is preferred to have) the ``Whole 
Disk'' partition. This partition references all of the sectors of the 
disk, and is used by the boot loader (either SILO, or Sun's).



and

It is also advised that the third partition should be of type ``Whole 
disk'' (type 5), and contain the entire disk (from the first cylinder 
to the last). This is simply a convention of Sun disk labels, and 
helps the |SILO| boot loader keep its bearings.


I also found this comment:

Make sure you create a ``Sun disk label'' on your boot disk. This is 
the only kind of partition scheme that the OpenBoot PROM understands, 
and so it's the only scheme from which you can boot. The /s/ key is 
used in |fdisk| to create Sun disk labels.




Nonetheless, even though I did not create a "Sun disk label" on my boot 
disk, or reserve the 3rd partition for the "Whole disk", everything 
seems to be working fine except for the speed issue. In other words, I'm 
not seeing the symptoms ascribed to my mistakes, so I'm wondering if 
that information is really accurate/up-to-date. However, since I am 
seeing speed issues, I wonder if these things are related. Is anyone 
confident enough that they are that you would recommend I rebuild the 
box? I'd rather not, but it wouldn't be catastrophic if I had to.


--
Kent

Эффективное постельное белье!

2003-10-02 Thread Подарки


Что бы вы не делали на нашем постельном белье,
это будет получаться гораздо лучше.
Отличный сон, гениальные дети и дьявольский секс...
Посетите магазин эффективного белья www.mypresent.ru 

Для отписки просьба воспользоваться специальной
формой на сайте. Спасибо за понимание!

Re: 2.6.0-test6

2003-10-02 Thread Paul

I am running on a sparc64 (ok, 2 of them), but I also have a sparc 32 (yes,
SMP)

Personally, I prefer 32 bit, all marketroid blathering to the contrary. If I
had a huge database or was simulating weather/gnabgig/etc, then it'd be
worthwhile. Otherwise, the overhead of 64 bit just isn't quite worth it to
me.

Do you have a pointer to known issues with 2.4.x and/or 2.6.x?
I'm not a kernel hacker, but it might just be time to try to muddle through
and at least identify issues to those who are probably better equipped than
I to fix.

- Original Message - 
From: "Ben Collins" <[EMAIL PROTECTED]>
To: "Paul" <[EMAIL PROTECTED]>
Cc: 
Sent: Thursday, October 02, 2003 5:48 PM
Subject: Re: 2.6.0-test6


> On Thu, Oct 02, 2003 at 04:54:07PM -0500, Paul wrote:
> > Ahh, but my goal is to have sparc32 running. I don't need 64 bt, and I'd
> > prefer not to have the overhead. I figure that sparc64 is a good place
to
> > start though. Not to mention the SS20 that I'm eventually going to have
> > running deb.
> >
> > Any tips on compiling sparc32 cleanly?
>
> I thought you were running on a sparc64. If that's so, you need to use
> egcs64 in order to compile kernels on woody, or gcc-3.3 in unstable,
> even though userspace is 32-bit.
>
> For sparc32, you will have lots of trouble getting 2.4, much less 2.6.x
> running on sparc32 at the moment. Mainly because no one has nailed down
> all the problems yet. 2.4.22 is pretty good, but still has issues on
> smp.
>
> -- 
> Debian - http://www.debian.org/
> Linux 1394 - http://www.linux1394.org/
> Subversion - http://subversion.tigris.org/
> WatchGuard - http://www.watchguard.com/
>
>
> -- 
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>

Re: 2.6.0-test6

On Thu, Oct 02, 2003 at 04:54:07PM -0500, Paul wrote:
> Ahh, but my goal is to have sparc32 running. I don't need 64 bt, and I'd
> prefer not to have the overhead. I figure that sparc64 is a good place to
> start though. Not to mention the SS20 that I'm eventually going to have
> running deb.
> 
> Any tips on compiling sparc32 cleanly?

I thought you were running on a sparc64. If that's so, you need to use
egcs64 in order to compile kernels on woody, or gcc-3.3 in unstable,
even though userspace is 32-bit.

For sparc32, you will have lots of trouble getting 2.4, much less 2.6.x
running on sparc32 at the moment. Mainly because no one has nailed down
all the problems yet. 2.4.22 is pretty good, but still has issues on
smp.

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/

Re: Are Sunblade 1000s slow?

On Thursday 02 October 2003 19:40, Ben Collins wrote:

> > /dev/scsi/host0/bus0/target1/lun0/part4 on /home type ext3 (rw)
> That's a rather ugly partition map :)
> Did you read the install docs where it says to not put a filesystem on
> partition number 3 and leave it as a Whole Disk to make OBP happy?

Here's my layout:

/dev/hda1 on / type ext3 (rw,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/hda4 on /usr type ext3 (rw)
/dev/hda5 on /home type ext3 (rw)

Is this correct? I had no significant experience in Sun disklabels and disk 
slices when I installed this machine, I hope to have did nothing negative :)

Have a great day,
Antonello

-- 
Antonello Iunco  - Natural Born Betatester
Cruising the Web on a (Modified) Debian-Powered Sun Ultra10

Re: Are Sunblade 1000s slow?

On Thursday 02 October 2003 19:57, Elie De Brauwer wrote:

> > ...and a ...umm...  I *think* 333MHz Ultra 5, but I could be off by a
> > few:
> > cpu : TI UltraSparc IIi
> > fpu : UltraSparc IIi integrated FPU
> > Cpu0Bogo: 719.25

Here is my Ultra10/333.

[EMAIL PROTECTED]:~$ cat /proc/cpuinfo
cpu : TI UltraSparc IIi (Sabre)
fpu : UltraSparc IIi integrated FPU
promlib : Version 3 Revision 15
prom: 3.15.2
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 665.19
Cpu0ClkTck  : 13d92d40
MMU Type: Spitfire

Your U5 CPU is a Blackbird, mine a Sabre. So you are probably posting a 
400-Mhz machine output. My machine is slower than yours according to specs. 
Your Blade runs slower than my lame Ultra10, it seems. I would like to buy a 
1500, hoping it will work correctly with Linux.

Bye,
Antonello

-- 
Antonello Iunco  - Natural Born Betatester
Cruising the Web on a (Modified) Debian-Powered Sun Ultra10

Re: Are Sunblade 1000s slow?

Kent West wrote:

Andrew Sharp wrote:

On Thu, Oct 02, 2003 at 07:57:13PM +0200, Elie De Brauwer wrote:

On Thursday 02 October 2003 19:29, Patrick Morris wrote:

Kent West wrote:

[EMAIL PROTECTED]:/home/snert> cat /proc/cpuinfo
cpu : TI UltraSparc III+ (Cheetah+)
fpu : UltraSparc III+ integrated FPU
promlib : Version 3 Revision 5
prom: 4.5.16
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 599.65

Er, that doesn't look right. Even if you have one of those
"hard-to-find"

(heh) sunblade1000's with a 600MHz cpu, then this should be more like
the other machines, ie., it should read Cpu0Bogo 1200. I'd say you
need to do some fixin on your kernel/compiler, or somebody else does.
Most likely, the version of gcc3 that was used has some problem
producing

code that works right on US3 cpus. This machine should be significantly
faster to use than your P3. Not only is the cpu faster, but the data
paths to memory and peripherals is a lot faster.

If you would like to send me your 1000 to play with for six months, I'm
sure I could figure out what's wrong. ~:^)

And oh yeah, don't waste any more of your valuable time trying to figure
anything out with glxgears.

Thanks for the response. I initially tried to get an "official" kernel
by following the links in the installation manual
(http://www.debian.org/releases/stable/sparc/ch-appendix.en.html#s-file-descs),
but those links
(http://http.us.debian.org/debian/dists/woody/main/disks-sparc/current/sparc64/tftpboot.img)
are dead. So I backed up a few levels from this address and then
spidered back down to
http://http.us.debian.org/debian/dists/woody/main/disks-sparc/current/sun4u/tftpboot.img,
but that image wouldn't boot on the Blade 1000. So I googled for
another image, and found the one at
http://auric.debian.org/~bcollins/disks-sparc/current/sparc64/ which
did boot and which is now working. What you're saying makes me think
my next thing to try will be to "apt-get install
kernel-image-2.4.21-sparc64". Thanks for the info! I'll post the
results of "cat /proc/cpuinfo" after the installation later today or
tomorrow.

Hmm, no difference.

Thu Oct 02 17:06:12
---
[EMAIL PROTECTED]:/home/snert> cat /proc/cpuinfo
cpu : TI UltraSparc III+ (Cheetah+)
fpu : UltraSparc III+ integrated FPU
promlib : Version 3 Revision 5
prom: 4.5.16
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 599.65
Cpu0ClkTck : 35a4e900
MMU Type: Cheetah+

Thu Oct 02 17:06:19
---
[EMAIL PROTECTED]:/home/snert> uname -a
Linux macs54 2.4.21 #1 Thu Aug 7 20:30:12 EDT 2003 sparc64 GNU/Linux

Re: 2.6.0-test6

2003-10-02 Thread Paul

Ahh, but my goal is to have sparc32 running. I don't need 64 bt, and I'd
prefer not to have the overhead. I figure that sparc64 is a good place to
start though. Not to mention the SS20 that I'm eventually going to have
running deb.

Any tips on compiling sparc32 cleanly?

Thanks

- Original Message - 
From: "Ben Collins" <[EMAIL PROTECTED]>
To: "Paul" <[EMAIL PROTECTED]>
Sent: Thursday, October 02, 2003 3:44 PM
Subject: Re: 2.6.0-test6

> On Thu, Oct 02, 2003 at 04:00:58PM -0500, Paul wrote:
> > 3.0r1 aka woody. It's a stock install.
> > I know I should upgrade the gcc to current, but haven't gotten there
yet.
>
> gcc3 in woody will compile broken kernels. Install egcs64 instead, and
> allow the build to use sparc64-linux-gcc as it should be doing.
>

Re: Are Sunblade 1000s slow?

Andrew Sharp wrote:

On Thu, Oct 02, 2003 at 07:57:13PM +0200, Elie De Brauwer wrote:

On Thursday 02 October 2003 19:29, Patrick Morris wrote:

Kent West wrote:

Er, that doesn't look right. Even if you have one of those "hard-to-find"
(heh) sunblade1000's with a 600MHz cpu, then this should be more like
the other machines, ie., it should read Cpu0Bogo 1200. I'd say you
need to do some fixin on your kernel/compiler, or somebody else does.
Most likely, the version of gcc3 that was used has some problem producing
code that works right on US3 cpus. This machine should be significantly
faster to use than your P3. Not only is the cpu faster, but the data
paths to memory and peripherals is a lot faster.

If you would like to send me your 1000 to play with for six months, I'm
sure I could figure out what's wrong. ~:^)

And oh yeah, don't waste any more of your valuable time trying to figure
anything out with glxgears.

Thanks for the response. I initially tried to get an "official" kernel
by following the links in the installation manual
(http://www.debian.org/releases/stable/sparc/ch-appendix.en.html#s-file-descs),
but those links
(http://http.us.debian.org/debian/dists/woody/main/disks-sparc/current/sparc64/tftpboot.img)
are dead. So I backed up a few levels from this address and then
spidered back down to
http://http.us.debian.org/debian/dists/woody/main/disks-sparc/current/sun4u/tftpboot.img,
but that image wouldn't boot on the Blade 1000. So I googled for another
image, and found the one at
http://auric.debian.org/~bcollins/disks-sparc/current/sparc64/ which did
boot and which is now working. What you're saying makes me think my next
thing to try will be to "apt-get install kernel-image-2.4.21-sparc64".
Thanks for the info! I'll post the results of "cat /proc/cpuinfo" after
the installation later today or tomorrow.

--
Kent

Re: Are Sunblade 1000s slow?

2003-10-02 Thread Andrew Sharp

On Thu, Oct 02, 2003 at 07:57:13PM +0200, Elie De Brauwer wrote:
> On Thursday 02 October 2003 19:29, Patrick Morris wrote:
> > Kent West wrote:
> > > [EMAIL PROTECTED]:/home/snert> cat /proc/cpuinfo
> > > cpu : TI UltraSparc III+ (Cheetah+)
> > > fpu : UltraSparc III+ integrated FPU
> > > promlib : Version 3 Revision 5
> > > prom: 4.5.16
> > > type: sun4u
> > > ncpus probed: 1
> > > ncpus active: 1
> > > Cpu0Bogo: 599.65

Er, that doesn't look right.  Even if you have one of those "hard-to-find"
(heh) sunblade1000's with a 600MHz cpu, then this should be more like
the other machines, ie., it should read Cpu0Bogo 1200.  I'd say you
need to do some fixin on your kernel/compiler, or somebody else does.
Most likely, the version of gcc3 that was used has some problem producing
code that works right on US3 cpus.  This machine should be significantly
faster to use than your P3.  Not only is the cpu faster, but the data
paths to memory and peripherals is a lot faster.

If you would like to send me your 1000 to play with for six months, I'm
sure I could figure out what's wrong.  ~:^)

And oh yeah, don't waste any more of your valuable time trying to figure
anything out with glxgears.

a

> > > Cpu0ClkTck  : 35a4e900
> > > MMU Type: Cheetah+
> >
> > Just for comparison, here's a dual-400 Ultra 2:
> >
> > cpu : TI UltraSparc II  (BlackBird)
> > fpu : UltraSparc II integrated FPU
> > promlib : Version 3 Revision 19
> > prom: 3.19.0
> > type: sun4u
> > ncpus probed: 2
> > ncpus active: 2
> > Cpu0Bogo: 799.53
> > Cpu0ClkTck  : 17d76a6d
> > Cpu1Bogo: 799.53
> > Cpu1ClkTck  : 17d76a6d
> > MMU Type: Spitfire
> > State:
> > CPU0:   online
> > CPU1:   online

> And while we're at it, this is a Sun Blade 100 with a 500 Mhz UltraSparc IIe.
> 
> [EMAIL PROTECTED]:~$ cat /proc/cpuinfo 
> cpu : TI UltraSparc IIe
> fpu : UltraSparc IIe integrated FPU
> promlib : Version 3 Revision 0
> prom: 4.0.45
> type: sun4u
> ncpus probed: 1
> ncpus active: 1
> Cpu0Bogo: 1002.70
> Cpu0ClkTck  : 1debe980
> MMU Type: Spitfire

Re: Mozilla Fails to start in Unstable

2003-10-02 Thread Martin

> You could install mozilla-snapshot from unstable which provides you with a 
> CVS 
> snapshot Build ID 2003061808 which works in a rather limited way (no PSM). 
> Another alternative is konqueror unless you're a lynx die hard ;-)
Bah!  Dillo for all your web browsing needs. :-)

But then agin - isn't that part of the appeal of Debian - you have the
choice.

Cheers,
 - Martin
 
-- 
Martin
[EMAIL PROTECTED]
"Seasons change, things come to pass"

Re: Mozilla Fails to start in Unstable


Elie De Brauwer wrote:


On Thursday 02 October 2003 19:26, Kent West wrote:
 


But after I dist-upgraded to
Unstable, Mozilla now fails to start.
 



You could install mozilla-snapshot from unstable which provides you with a CVS 
snapshot Build ID 2003061808 which works in a rather limited way (no PSM). 
Another alternative is konqueror unless you're a lynx die hard ;-)
 

Thanks for the ideas, but nah; I'm just running Moz remotely from my 
x86-based Debian box until the problems get fixed.


--
Kent

Re: Mozilla Fails to start in Unstable

2003-10-02 Thread Elie De Brauwer

On Thursday 02 October 2003 19:26, Kent West wrote:
> Kent West wrote:
> >But after I dist-upgraded to
> >Unstable, Mozilla now fails to start.
>
> Elie De Brauwer wrote:
> >Don't even start about it, here mozilla is broken (along with everything
> > that depends on it like galeon and firebird) for about a month now. The
> > bugreport i submitted is 207743 without much succes by now.  Maybe the
> > problem is related.
> >
> >Greetings
>
> Ah, thanks! They don't call it "unstable" for nuthin'.  :-)  Your reply
> just saved me from making another post about Galeon.
>

You could install mozilla-snapshot from unstable which provides you with a CVS 
snapshot Build ID 2003061808 which works in a rather limited way (no PSM). 
Another alternative is konqueror unless you're a lynx die hard ;-)


-- 
http://www.de-brauwer.be

Re: Are Sunblade 1000s slow?

2003-10-02 Thread Elie De Brauwer

On Thursday 02 October 2003 19:29, Patrick Morris wrote:
> Kent West wrote:
> > [EMAIL PROTECTED]:/home/snert> cat /proc/cpuinfo
> > cpu : TI UltraSparc III+ (Cheetah+)
> > fpu : UltraSparc III+ integrated FPU
> > promlib : Version 3 Revision 5
> > prom: 4.5.16
> > type: sun4u
> > ncpus probed: 1
> > ncpus active: 1
> > Cpu0Bogo: 599.65
> > Cpu0ClkTck  : 35a4e900
> > MMU Type: Cheetah+
>
> Just for comparison, here's a dual-400 Ultra 2:
>
> cpu : TI UltraSparc II  (BlackBird)
> fpu : UltraSparc II integrated FPU
> promlib : Version 3 Revision 19
> prom: 3.19.0
> type: sun4u
> ncpus probed: 2
> ncpus active: 2
> Cpu0Bogo: 799.53
> Cpu0ClkTck  : 17d76a6d
> Cpu1Bogo: 799.53
> Cpu1ClkTck  : 17d76a6d
> MMU Type: Spitfire
> State:
> CPU0:   online
> CPU1:   online
>
> ...and a ...umm...  I *think* 333MHz Ultra 5, but I could be off by a few:
>
> cpu : TI UltraSparc IIi
> fpu : UltraSparc IIi integrated FPU
> promlib : Version 3 Revision 25
> prom: 3.25.1
> type: sun4u
> ncpus probed: 1
> ncpus active: 1
> Cpu0Bogo: 719.25
> Cpu0ClkTck  : 15752a00
> MMU Type: Spitfire

And while we're at it, this is a Sun Blade 100 with a 500 Mhz UltraSparc IIe.

[EMAIL PROTECTED]:~$ cat /proc/cpuinfo 
cpu : TI UltraSparc IIe
fpu : UltraSparc IIe integrated FPU
promlib : Version 3 Revision 0
prom: 4.0.45
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 1002.70
Cpu0ClkTck  : 1debe980
MMU Type: Spitfire


-- 
http://www.de-brauwer.be

Re: Are Sunblade 1000s slow?

> I suspected the disk, but hdparm is the only tool I know of for dealing 
> with that, but I believe that's only for IDE drives, and this box has scsi:

SCSI should be good.

> [EMAIL PROTECTED]:/home/snert> mount
> /dev/scsi/host0/bus0/target1/lun0/part1 on / type ext3 
> (rw,errors=remount-ro)
> proc on /proc type proc (rw)
> devpts on /dev/pts type devpts (rw,gid=5,mode=620)
> /dev/scsi/host0/bus0/target1/lun0/part2 on /usr type ext3 (rw)
> /dev/scsi/host0/bus0/target1/lun0/part3 on /var type ext3 (rw)
> /dev/scsi/host0/bus0/target1/lun0/part6 on /tmp type ext3 (rw)
> /dev/scsi/host0/bus0/target1/lun0/part7 on /spare1 type ext3 (rw)
> /dev/scsi/host0/bus0/target1/lun0/part8 on /spare2 type ext3 (rw)
> /dev/scsi/host0/bus0/target1/lun0/part4 on /home type ext3 (rw)

That's a rather ugly partition map :)

Did you read the install docs where it says to not put a filesystem on
partition number 3 and leave it as a Whole Disk to make OBP happy?

I honestly don't know what your performance problems are asttributed to.
Maybe you are expecting too much for your money :) Sun's hardware is
expensive, and generally isn't as fast in some areas as PC's. They are,
however, more reliable, and generally better for scientific computing
(e.g. they get better FPU performance). Not to mention the ability to do
64-bit computing without hassle.

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/

Re: Are Sunblade 1000s slow?

2003-10-02 Thread Amilcar Meneses


I'm goint to try install for compare with you...

Amilcar

On Thu, 2 Oct 2003, Kent West wrote:

> Amilcar Meneses wrote:
>
> >How do you install debian?
> >
> >Amilcar
> >
> >
> >
> We've got a Sun server that runs a tftp server and a dhcp server; so I
> configured it to provide a linux rescue/root/boot/install kernel from
> http://auric.debian.org/~bcollins/disks-sparc/current/sparc64/ to the
> Sunblade 1000's hardware address, and then did a "boot net" from the
> OpenFirmWare prompt on the Sunblade 1000; it picked up the bcollins
> kernel and started the Debian install. I pulled the remainder of Debian
> from the internet via http when the installer got to that point. Then
> after Stable was finished installing, I changed my /etc/apt/sources.list
> to point to Unstable and did a dist-upgrade.
>
> Is that what you were asking?
>
> --
> Kent
>
>
>
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
>

Re: Are Sunblade 1000s slow?

2003-10-02 Thread Patrick Morris


Kent West wrote:


[EMAIL PROTECTED]:/home/snert> cat /proc/cpuinfo
cpu : TI UltraSparc III+ (Cheetah+)
fpu : UltraSparc III+ integrated FPU
promlib : Version 3 Revision 5
prom: 4.5.16
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 599.65
Cpu0ClkTck  : 35a4e900
MMU Type: Cheetah+


Just for comparison, here's a dual-400 Ultra 2:

cpu : TI UltraSparc II  (BlackBird)
fpu : UltraSparc II integrated FPU
promlib : Version 3 Revision 19
prom: 3.19.0
type: sun4u
ncpus probed: 2
ncpus active: 2
Cpu0Bogo: 799.53
Cpu0ClkTck  : 17d76a6d
Cpu1Bogo: 799.53
Cpu1ClkTck  : 17d76a6d
MMU Type: Spitfire
State:
CPU0:   online
CPU1:   online

...and a ...umm...  I *think* 333MHz Ultra 5, but I could be off by a few:

cpu : TI UltraSparc IIi
fpu : UltraSparc IIi integrated FPU
promlib : Version 3 Revision 25
prom: 3.25.1
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 719.25
Cpu0ClkTck  : 15752a00
MMU Type: Spitfire

Re: Mozilla Fails to start in Unstable


Kent West wrote:


But after I dist-upgraded to
Unstable, Mozilla now fails to start.
 




Elie De Brauwer wrote:

Don't even start about it, here mozilla is broken (along with everything that 
depends on it like galeon and firebird) for about a month now. The bugreport 
i submitted is 207743 without much succes by now.  Maybe the problem is 
related.


Greetings
 



Ah, thanks! They don't call it "unstable" for nuthin'.  :-)  Your reply 
just saved me from making another post about Galeon.


--
Kent

Re: Mozilla Fails to start in Unstable


Ben Collins wrote:


On Thu, Oct 02, 2003 at 11:10:47AM -0500, Kent West wrote:
 

But after I dist-upgraded to 
Unstable, Mozilla now fails to start.
   


Known problem with mozilla. Try to get the 1.4-2 packages. Maybe they
are in testing.

 


Ah, thanks! They don't call it "unstable" for nuthin'.  :-)

--
Kent

Re: Are Sunblade 1000s slow?


Ben Collins wrote:


On Thu, Oct 02, 2003 at 10:42:59AM -0500, Kent West wrote:
 


Kent West wrote:

   

This year, I've gotten the opportunity to put Debian on a Sunblade 
1000 (with 1GB RAM - whoo-hoo!). However, now that it's running 
Debian, it seems awfully sluggish.
 


Another example: running "glxgears" only shows a frame rate of 26 to 44 FPS.
   



GL isn't a good comparison, since it relies on video hardware, not the
CPU (unless of course you are running it as software-GL).
 

Yeah, I wasn't too reliant on the GL numbers, but still 44 seems awfully 
slow to me. I've tried using both the ATI driver and the fbdev driver 
for xfree86, and I've tried both the with framebuffer and without 
option; the only thing that changes is that it locks up the machine hard 
forcing a power cycle to recover.


But as mentioned in previous posts, even console-based activities seem 
sluggish. Do this info look right?


[EMAIL PROTECTED]:/home/snert> cat /proc/cpuinfo
cpu : TI UltraSparc III+ (Cheetah+)
fpu : UltraSparc III+ integrated FPU
promlib : Version 3 Revision 5
prom: 4.5.16
type: sun4u
ncpus probed: 1
ncpus active: 1
Cpu0Bogo: 599.65
Cpu0ClkTck  : 35a4e900
MMU Type: Cheetah+

Re: Are Sunblade 1000s slow?


Ben Collins wrote:


On Thu, Oct 02, 2003 at 10:42:59AM -0500, Kent West wrote:
 


Kent West wrote:

   

This year, I've gotten the opportunity to put Debian on a Sunblade 
1000 (with 1GB RAM - whoo-hoo!). However, now that it's running 
Debian, it seems awfully sluggish.
 


Another example: running "glxgears" only shows a frame rate of 26 to 44 FPS.
   



GL isn't a good comparison, since it relies on video hardware, not the
CPU (unless of course you are running it as software-GL).

I have a Blade100, and I don't see it as being very slow. Maybe you are
comparing the wrong things. What sort of disk does the Blade 1000 have?
Is it IDE or scsi? Disk access depends more on the disk and disk
controller, and less to do with CPU.

 

I suspected the disk, but hdparm is the only tool I know of for dealing 
with that, but I believe that's only for IDE drives, and this box has scsi:


[EMAIL PROTECTED]:/home/snert> mount
/dev/scsi/host0/bus0/target1/lun0/part1 on / type ext3 
(rw,errors=remount-ro)

proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/scsi/host0/bus0/target1/lun0/part2 on /usr type ext3 (rw)
/dev/scsi/host0/bus0/target1/lun0/part3 on /var type ext3 (rw)
/dev/scsi/host0/bus0/target1/lun0/part6 on /tmp type ext3 (rw)
/dev/scsi/host0/bus0/target1/lun0/part7 on /spare1 type ext3 (rw)
/dev/scsi/host0/bus0/target1/lun0/part8 on /spare2 type ext3 (rw)
/dev/scsi/host0/bus0/target1/lun0/part4 on /home type ext3 (rw)

Re: Mozilla Fails to start in Unstable

2003-10-02 Thread Elie De Brauwer

Don't even start about it, here mozilla is broken (along with everything that 
depends on it like galeon and firebird) for about a month now. The bugreport 
i submitted is 207743 without much succes by now.  Maybe the problem is 
related.

Greetings

> I've got a Sunblade 1000 with a fresh install of Debian. I originally
> installed Stable, and Mozilla ran. But after I dist-upgraded to
> Unstable, Mozilla now fails to start.
>
> I tried moving my ~/.mozilla directory out of the way; I tried
> purging/reinstalling Mozilla; no difference.
>
> When I try starting Moz from a terminal window, I get:
>
> Thu Oct 02   10:37:35
> ---
> [EMAIL PROTECTED]:/home/snert> mozilla
> /dev/dsp: No such file or directory
>
> Thu Oct 02   10:52:28
> ---
> [EMAIL PROTECTED]:/home/snert> ps ax | grep moz
> 26894 pts/0S  0:00 grep moz
>
>
> When I originally installed Moz, when it asked what sound daemon to use,
> I believe I chose "auto". Now when I run "dpkg-reconfigure mozilla",
> wanting to set it to none, it "thinks" for 20 or 30 seconds, and then
> nothing happens:
>
> Thu Oct 02   10:52:38
> ---
> [EMAIL PROTECTED]:/home/snert> sudo dpkg-reconfigure mozilla
> Password:
>
> Thu Oct 02   10:54:09
> ---
> [EMAIL PROTECTED]:/home/snert>
>
>
> Any clues?
>
> Thanks!
>
> --
> Kent

-- 
http://www.de-brauwer.be

Re: Mozilla Fails to start in Unstable

On Thu, Oct 02, 2003 at 11:10:47AM -0500, Kent West wrote:
> I've got a Sunblade 1000 with a fresh install of Debian. I originally 
> installed Stable, and Mozilla ran. But after I dist-upgraded to 
> Unstable, Mozilla now fails to start.
> 
> I tried moving my ~/.mozilla directory out of the way; I tried 
> purging/reinstalling Mozilla; no difference.
> 
> When I try starting Moz from a terminal window, I get:

Known problem with mozilla. Try to get the 1.4-2 packages. Maybe they
are in testing.

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/

Re: Are Sunblade 1000s slow?

On Thu, Oct 02, 2003 at 10:42:59AM -0500, Kent West wrote:
> Kent West wrote:
> 
> >This year, I've gotten the opportunity to put Debian on a Sunblade 
> >1000 (with 1GB RAM - whoo-hoo!). However, now that it's running 
> >Debian, it seems awfully sluggish.
> 
> Another example: running "glxgears" only shows a frame rate of 26 to 44 FPS.

GL isn't a good comparison, since it relies on video hardware, not the
CPU (unless of course you are running it as software-GL).

I have a Blade100, and I don't see it as being very slow. Maybe you are
comparing the wrong things. What sort of disk does the Blade 1000 have?
Is it IDE or scsi? Disk access depends more on the disk and disk
controller, and less to do with CPU.

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/

Mozilla Fails to start in Unstable

I've got a Sunblade 1000 with a fresh install of Debian. I originally 
installed Stable, and Mozilla ran. But after I dist-upgraded to 
Unstable, Mozilla now fails to start.


I tried moving my ~/.mozilla directory out of the way; I tried 
purging/reinstalling Mozilla; no difference.


When I try starting Moz from a terminal window, I get:

Thu Oct 02   10:37:35
---
[EMAIL PROTECTED]:/home/snert> mozilla
/dev/dsp: No such file or directory

Thu Oct 02   10:52:28
---
[EMAIL PROTECTED]:/home/snert> ps ax | grep moz
26894 pts/0S  0:00 grep moz


When I originally installed Moz, when it asked what sound daemon to use, 
I believe I chose "auto". Now when I run "dpkg-reconfigure mozilla", 
wanting to set it to none, it "thinks" for 20 or 30 seconds, and then 
nothing happens:


Thu Oct 02   10:52:38
---
[EMAIL PROTECTED]:/home/snert> sudo dpkg-reconfigure mozilla
Password:

Thu Oct 02   10:54:09
---
[EMAIL PROTECTED]:/home/snert>


Any clues?

Thanks!

--
Kent

Re: 2.6.0-test6

2003-10-02 Thread Magosányi Árpád

A levelezÅm azt hiszi, hogy Ben Collins a kÃ¶vetkezÅeket Ãrta:

> You never said what ver of Debian-sparc you are running. If you are
> using gcc-3, I hope you are using atleast 3.3.2.

Do we have a list of known to work and known not to work
packages? Especially of those which are needed to compile the kernel.
As I am also experimenting with 2.6.0, it would be helpful to concentrate
only the problems of the kernel itself, as apparently it have
a lot of them anyway.

It is also would be good to have a list of known problems, and
how to report the new ones. I have tried to report to davem and to the
netfilter list various problems in the way it is described in the
kernel source, and there was no answer.

My list of problems (I am on sparc64):
'iptables -N test' hangs the kernel completely.
the canonical /dev/console does not seem to work on my headless machine
slapd answers _very_ slowly (may be unrelated to kernel problems)

Any hints appreciated, I can send detailed bug reports, do
experiments, etc on request.

-- 
GNU GPL: csak tiszta forrÃ¡sbÃ³l

Re: Are Sunblade 1000s slow?


Kent West wrote:

This year, I've gotten the opportunity to put Debian on a Sunblade 
1000 (with 1GB RAM - whoo-hoo!). However, now that it's running 
Debian, it seems awfully sluggish.


Another example: running "glxgears" only shows a frame rate of 26 to 44 FPS.

Re: Are Sunblade 1000s slow?


Amilcar Meneses wrote:


How do you install debian?

Amilcar

 

We've got a Sun server that runs a tftp server and a dhcp server; so I 
configured it to provide a linux rescue/root/boot/install kernel from
http://auric.debian.org/~bcollins/disks-sparc/current/sparc64/ to the 
Sunblade 1000's hardware address, and then did a "boot net" from the 
OpenFirmWare prompt on the Sunblade 1000; it picked up the bcollins 
kernel and started the Debian install. I pulled the remainder of Debian 
from the internet via http when the installer got to that point. Then 
after Stable was finished installing, I changed my /etc/apt/sources.list 
to point to Unstable and did a dist-upgrade.


Is that what you were asking?

--
Kent

capstan wheel

2003-10-02 Thread Piotr Madrzyk




Hi,
Could you let me know if you 
have capstan wheels for Archive 2150s tape drive.
Peter

Re: Are Sunblade 1000s slow?

2003-10-02 Thread Amilcar Meneses


How do you install debian?

Amilcar

On Thu, 2 Oct 2003, Kent West wrote:

> I've typically run Debian on x86 hardware and on Mac G3/G4s, and love
> Debian.
>
> A couple of years ago I started dabbling with Sun machines, specifically
> Sunblade 100s running Solaris 8 (having never had any Sun/Solaris/other
> unix (other than Linux) experience before). Suffice it to say I was not
> impressed with Solaris. So last summer I put Debian on one of the
> Sunblade 100s for a week or two before it had to be converted back to
> Solaris. I didn't have long to play with it, but I liked Debian better
> than Solaris on the box. But that was last year, and I really don't have
> much memory of it (except that the nfs connection to a Sun server get
> hanging the box).
>
> This year, I've gotten the opportunity to put Debian on a Sunblade 1000
> (with 1GB RAM - whoo-hoo!). However, now that it's running Debian, it
> seems awfully sluggish. Of course, it seemed somewhat sluggish under
> Solaris 8 also, but I was hoping Debian would make a difference.
> However, Debian feels like it's running on a 200MHz Pentium II box. My
> 933MHz PIII feels like it runs circles around this Sunblade 1000.
>
> So are Sunblade 1000s just inherently slower than the average x86
> machines, or is there some tweak that I'm missing, or what? It seems
> slower in both console and X. Like for example, I'll do an "apt-get
> upgrade", and the downloading of the packages from the net seems about
> the same on the two architectures, but when apt starts unpacking and
> configuring the packages, it just seems slower on the Sun box. Another
> example is when I run "dpkg-reconfigure xserver-xfree86"; it just seems
> to take 15 or 20 seconds of "thinking"  between the command and the
> appearance of the ncurses dialog on the Sun, whereas the same command on
> my PIII comes up almost instantly. Or when I copy and paste this email
> message into a nano session in an xterm window on the Sun box, it draws
> each line slow enough that I can almost read it as it's being pasted (to
> be fair, I'm a fast reader).
>
> Thanks for any insight!
>
> --
> Kent
>
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
>
>

Are Sunblade 1000s slow?

I've typically run Debian on x86 hardware and on Mac G3/G4s, and love 
Debian.


A couple of years ago I started dabbling with Sun machines, specifically 
Sunblade 100s running Solaris 8 (having never had any Sun/Solaris/other 
unix (other than Linux) experience before). Suffice it to say I was not 
impressed with Solaris. So last summer I put Debian on one of the 
Sunblade 100s for a week or two before it had to be converted back to 
Solaris. I didn't have long to play with it, but I liked Debian better 
than Solaris on the box. But that was last year, and I really don't have 
much memory of it (except that the nfs connection to a Sun server get 
hanging the box).


This year, I've gotten the opportunity to put Debian on a Sunblade 1000 
(with 1GB RAM - whoo-hoo!). However, now that it's running Debian, it 
seems awfully sluggish. Of course, it seemed somewhat sluggish under 
Solaris 8 also, but I was hoping Debian would make a difference. 
However, Debian feels like it's running on a 200MHz Pentium II box. My 
933MHz PIII feels like it runs circles around this Sunblade 1000.


So are Sunblade 1000s just inherently slower than the average x86 
machines, or is there some tweak that I'm missing, or what? It seems 
slower in both console and X. Like for example, I'll do an "apt-get 
upgrade", and the downloading of the packages from the net seems about 
the same on the two architectures, but when apt starts unpacking and 
configuring the packages, it just seems slower on the Sun box. Another 
example is when I run "dpkg-reconfigure xserver-xfree86"; it just seems 
to take 15 or 20 seconds of "thinking"  between the command and the 
appearance of the ncurses dialog on the Sun, whereas the same command on 
my PIII comes up almost instantly. Or when I copy and paste this email 
message into a nano session in an xterm window on the Sun box, it draws 
each line slow enough that I can almost read it as it's being pasted (to 
be fair, I'm a fast reader).


Thanks for any insight!

--
Kent

Re: [debian-sparc] Update on mplayer support for Linux/SPARC

On Thu, 2 Oct 2003 14:27:48 +0200 (CEST)
Erwann Abalea <[EMAIL PROTECTED]> wrote:

> BTW, without mlib, I can get a pretty good video experience on my 300 MHz
> Ultra2, but I have to allow the player to drop frames to get audio/video
> sync. When I'm only watching the video and nothing else runs, I sometimes

I solved many AV sync issues by using SDL audio and video output
drivers. However, on my Ultra 10, maybe because of the SLOW ide
controller and of the huge windowing system that I use (KDE 3, dual
display) it's quite impossible to watch a movie with a human-compliant
frame rate (at least on the PGX) ;). With mlib and the Creator board
instead I'm able to watch a movie (windowed, not in fullscreen mode)
with frame dropping disabled. I do have a SCSI controller, and an
external SCSI box for 6 SCA disks (think it's called an UniPack), but
the disks are old and their capacity is limited. I think the lack of
decent IDE busmastering on the Ultra10 is the main reason for my system
being slow as a little snail ;).

Now I'm trying to convince my employer to buy a new Sun workstation to
replace my x86 PC. I would to run Debian on it (side by side with
Solaris), so I am wondering if someone the list might give me some
advice on the most Penguin-compatible and valuable workstation to buy
(at a reasonable price, of course) ;). Any suggestion is appreciated ;)

Bye,
Antonello

-- 
Antonello <[EMAIL PROTECTED]> ICQ 8442040-MSN [EMAIL PROTECTED] - phone: ask :P

There will be no consolation prize / this time the bone is broken clean
no baptism, no reprise / and no sweet taste of victory [Jann Arden]

Re: 2.6.0-test6

> I'm using gcc3 (ln -s) and it's stock sparc64 beyond that. The sparc port 
> seems to have issues with gcc/register clobbers, I haven't dug up the patch 
> for that. 

You never said what ver of Debian-sparc you are running. If you are
using gcc-3, I hope you are using atleast 3.3.2.

> Biggest complaint: Small allowable kernel size, about 3.5 meg.~1.42Meg 
> gzipped. I'm debating trying to delve into the mysteries of OBP/SILO/Boottime 
> mem management and see if there is some kind of workaround.

When you build your kernel, do "make image" and use the
arch/sparc64/boot/image file to boot from. It gets stripped of all the
normal symbols that just bloat the image.

-- 
Debian - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/

2.6.0-test6

2003-10-02 Thread Paul




Well, I'm new here so a quick personal sparc 
rundown: 
SS20 Dual SM62's,
Ultra 1 170, Creator3d series 2, 2 
qfe's
Ultra 10 440
 
Anyway, I've got 2.6.0-test6 to compile and boot, 
and I thought I might pass along a couple of wierdnesses I stumbled across 
(please note that it booted, I haven't quite yet updated the modutils, etc, 
etc)
 
Wierdest thing: If I try to boot a vmlinux, I get a 
hard hang. I have to physically flip the power switch to recover. BUT, if I gzip 
and rename, it boots. 
 
gzip vmlinux, mv vmlinux.gz vmlinuz (not necessary, 
but prettier)
 
SILO boot: 1/boot/vmlinuz 
 
This works, but skip the above steps and booting 
1/boot/vmlinux hard hangs.
 
Another wierd thing is probably something simple 
I'm missing. I finally get to a login prompt and the keyboard is 'remapped'. 
asdf comes out as 1234, even get the shifted symbols.
 
I'm using gcc3 (ln -s) and it's stock sparc64 
beyond that. The sparc port seems to have issues with gcc/register clobbers, I 
haven't dug up the patch for that. 
 
Biggest complaint: Small allowable kernel size, 
about 3.5 meg.~1.42Meg gzipped. I'm debating trying to delve into the mysteries 
of OBP/SILO/Boottime mem management and see if there is some kind of 
workaround.
 
Anyway, just wanted to ...oh heck, I'll admit it: 
Crow a little! 
 
(Now, if I could just stumble across a real cheap 
bootable SCSI card for me U10, now THEN I would be a happy 
camper)

Re: [debian-sparc] Update on mplayer support for Linux/SPARC

2003-10-02 Thread Erwann Abalea

On Thu, 2 Oct 2003, Antonello wrote:

> Where did you find the source deb-src? do you mean the official Debian
> archive?

Sorry, that was an older version (0.90-rc5). I downloaded the 1.0pre1
version from the MPlayer home page, there's a debian/ directory in the
archive, and a 'dpkg-buildpackage' does the rest. The resulting debian
package is named mplayer_0.90cvs_sparc.deb, though.

> > How can I say if mplayer was compiled with mlib support? Will a 'ldd
> > /usr/bin/mplayer' show something, or is something loaded at runtime?
>
> If compiled with mlib, mplayer outputs in console a message that sounds
> like: "using mlib for colorspace transform"

I'll test it when I'll be back home.

> If you haven't enabled mlib yet, I advice you to do it, since the
> performance boost (at no additional cost) is remarkable. I posted a
> pico-howto on this subject some time ago (with the precious help of Mr.
> Miller), you can find it in the mailing list archives.

I found your pico-howto, and I'll test it. I'm downloading the libmlib
packages, I'll test them later. Thanks.

BTW, without mlib, I can get a pretty good video experience on my 300 MHz
Ultra2, but I have to allow the player to drop frames to get audio/video
sync. When I'm only watching the video and nothing else runs, I sometimes
can notice that frames were dropped; the margin is very low, I can't even
run a simple 'top' in a console, or I loose too much frames. The film was
accessed by NFS, over a 10 MBps link (yes, between my workstation and my
fileserver, there's a 10 MBps hub somewhere, I have to find a cheap and
small 100 MBps one).

-- 
Erwann ABALEA <[EMAIL PROTECTED]> - RSA PGP Key ID: 0x2D0EABD5
-
Le netétiquette n'est qu'une vaste fumisterie,il faut de l'argent pour
fonctionner,à force,en France de refuser tout rapport sain avec
l'argent,l'on riqsque de tuer ce nouvel outil.
-+- AA in: Guide du Neuneu d'Usenet - Le netétiquette du riche -+-

Re: debian/sparc serial console

2003-10-02 Thread Mauricio


At 18:54 +0200 10/1/03, Antonello wrote:

On Wed, 1 Oct 2003 18:22:04 +0100
Emmanuel Kasper <[EMAIL PROTECTED]> wrote:


 > Um, with my SS10, ROM 2.25, this doesn't happen. I have to explicitely
 > set serial I/O if I want that. Simply unplugging keyboard (which works
 > nicely on another SS10 with an older ROM version) doesn't help here...
 I also have an Ultra10, and  when you boot, the PROM expects a

 > console on the serial port if it doesn't detect the sun keyboard.

Emmanuel,
The SparcStation 10 is not an Ultra ;)
However, my Ultra2 (with Woody) is completely headless and behaves the
same way that you describe.. even the "keyboard not found, using tty for
i/o" appears on the serial console (9600, N, 1, no flow control).


	I would also add that in my SS20 (ROM 2.25R) and in my SS1+, 
if I do not have a serial console connected to its serial port when I 
boot them keyboard-less, they refuse to write to the serial console 
once it is connected.  I do not know if the events are related or it 
is just a freak coincidence, but...

Re: Broken keyboard in XFree86

2003-10-02 Thread Steve Pacenka

On Wed, 2003-10-01 at 09:42, Thurmond, Phillip wrote:
> I have a newly installed debian (unstable) system.  So far, I can’t
> get X to work.  The mouse and display work fine, however the keyboard
> is completely unresponsive.  The machine doesn’t even respond to
> ctrl-alt-del, stop-a, or ctrl-alt-bspc.  If I ssh in and kill the X
> server, I’m dropped to a console at which point the keyboard works
> fine again.  I’ve seen many messages about problems with keymaps, but
> none where the keyboard just didn’t work at all.  If anyone could help
> I’d appreciate it.
> 


> Here is the relevant section from my XF86Config-4.  I’ve tried it with
> and without the “Option” lines.
> 
>  
> 
> Section "InputDevice"
> 
> Identifier  "Keyboard0"
> 
> Driver  "keyboard"
> 
> Option "CoreKeyboard"
> 
> Option  "XkbRules" "sun"
> 
> Option "XkbModel" "type6"
> 
> Option "XkbLayout" "us"
> 
> EndSection

Just to make sure, does this file also refer to that keyboard spec in
the ServerLayout section:

Section "ServerLayout"
...
...
InputDevice "Keyboard0"
...
...
EndSection


And have you tried other keyboard types, like "type5"?  I'm not sure
what type my U10's keyboard is, but it's newer than Type 5s I own and I
still use the type5 model in XF86Config-4.  I don't remember having a
type6 choice (in this case 4.2.1-6) during setup.

-- good luck, SP

Re: [debian-sparc] Update on mplayer support for Linux/SPARC