Re: unofficial 2.4.21-pre1 sparc32 kernel debs

2002-12-12 Thread Christian Jönsson
On Wed, Dec 11, 2002 at 10:18:43PM -0500, Rob Radez wrote:
 Unofficial 2.4.21-pre1 sparc32 kernel debs have been uploaded to
 http://osinvestor.com/sparc/debs/ .  Running UP here, but I can make no
 guarantees.  Also, does anyone even care about this now that official
 debian 2.4 sparc debs are available?

I must have missed that, where is that official site?

Cheers,

/ChJ



U5, 2.4.20, sunhme

2002-12-12 Thread Antonio Prioglio
Further to my previious message, I patched sunhme.c with the 21pre1 
correction but I can observe no difference, i.e. RX only not TX.

Connection is to a 3com 3c16471.

=
eth0  Link encap:Ethernet  HWaddr 08:00:20:B6:22:F1  
  UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
  RX packets:3 errors:0 dropped:0 overruns:0 frame:0
  TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
  collisions:0 txqueuelen:100 
  RX bytes:623 (623.0 b)  TX bytes:0 (0.0 b)
  Interrupt:224 Base address:0xb000 


Any suggestion in the cosideration that 2.4.19 is unable to read/write to 
the hw clock? (2.4.18 is even worse!)

Is my experience unique?

-- 
Regards,
Antonio Prioglio [EMAIL PROTECTED]
http://www.cbtcentre.co.uk

Cognitive Behavioural Therapy Centre - London



Re: 2.4.20 kernel link failure: undefined reference to `_mcount'

2002-12-12 Thread Dave Love
Holt Sorenson [EMAIL PROTECTED] writes:

 I'm attempting to compile kernel 2.4.20 on a sun ultra 2 using 
 gcc-3.0 (3.0.4) from debian gnu/linux sparc woody 3.0r0.

Don't.  Use `egcs64' -- see archives of this list.  Does the procedure
in the kernel section of the Debian FAQ not work?  If it doesn't,
please make a bug report.

[What's the advantage of that kernel over the latest Debian package
anyway?]

 The linux kernel faq 
 (http://www.wsinf.edu.pl/doc/doc-linux-html/FAQ/Linux-FAQ-8.html)
 suggests that the linker is using a static libc for linking.

I don't think that's a kernel FAQ, despite the name.



Re: 2.4.20 kernel link failure: undefined reference to `_mcount'

2002-12-12 Thread Roy Bixler
On Thu, Dec 12, 2002 at 05:17:29PM +, Dave Love wrote:
 Holt Sorenson [EMAIL PROTECTED] writes:
 
  I'm attempting to compile kernel 2.4.20 on a sun ultra 2 using 
  gcc-3.0 (3.0.4) from debian gnu/linux sparc woody 3.0r0.
 
 Don't.  Use `egcs64' -- see archives of this list.  Does the procedure
 in the kernel section of the Debian FAQ not work?  If it doesn't,
 please make a bug report.

I have successfully used the latest gcc-3.2.1 builds to compile kernel
2.4.20 on a Sunblade 100 (Ultra IIe processor).  At least, the kernel
has been up for more than 7 days without a problem.

-- 
Roy Bixler [EMAIL PROTECTED]
The University of Chicago Press



Re: mkisofs and silo

2002-12-12 Thread Joerg Jaspert
Ben Collins [EMAIL PROTECTED] writes:

  I'll NMU with the patch if you ignore this.
 Don't NMU any of my packages without the normal NMU procedure.
 Don't remove essential patches for no other reason than spring cleaning.

I talked with upstream again.
Patch will stay for 2.0, but in 2.1alpha he adds code against that/that
breaks with it. Something like this. Then its deleted.
(Dont ask me about license that part will get.)
Then linux sparc could stay with last major release of mkisofs or adjust
silo. (His words). 

 The patch hurts nothing and breaks nothing, but provides a need part
 of the sparc port.

Then it hurts/breaks. But someone could then built mkisofs-sparc
packages if one wants (old major release) or do something completly different.
Debian Package should stay as close as possible to official version,
so it wont have it too.
(I dont want to lose Upstream support for every Debian User just
because of a patch wich is not necessary if the real buggy thing is fixed.)

Ah, and it still applies, dont NMU one of my packages without following
the normal procedure for an NMU. :)

BTW: I announced at 21 Nov 2002 that this patch is near his dead. Noone
cared about so It doesnt look very important to me if there is no
response from anyone.

BTW2: I dont want to flame anyone anywhere, just in case someone thinks that.

-- 
bye Joerg
A.D. 1492:
Christopher Columbus arrives in what he believes to be India, but
which RMS informs him is actually GNU/India.


pgpZVOF4O4I2K.pgp
Description: PGP signature


Re: mkisofs and silo

2002-12-12 Thread Ben Collins
On Fri, Dec 13, 2002 at 12:07:35AM +0100, Joerg Jaspert wrote:
 Ben Collins [EMAIL PROTECTED] writes:
 
   I'll NMU with the patch if you ignore this.
  Don't NMU any of my packages without the normal NMU procedure.
  Don't remove essential patches for no other reason than spring cleaning.
 
 I talked with upstream again.
 Patch will stay for 2.0, but in 2.1alpha he adds code against that/that
 breaks with it. Something like this. Then its deleted.
 (Dont ask me about license that part will get.)
 Then linux sparc could stay with last major release of mkisofs or adjust
 silo. (His words). 

He is being really childish if he is intentionally making things break
for no good reason.

Worse comes to worse, I'll redo the patch like I did before.


Don't let an overzealous upstream tell us what we can and cannot do with
free source. We are allowed to change it, and do whatever we want with
it. If he makes it so we can't, then it isn't free anymore and the whole
issue is moot, since we need to find a isofs creation tool that is free.

Tell upstream to stay out of Debian business.

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



Re: 2.4.20 kernel link failure: undefined reference to `_mcount'

2002-12-12 Thread Holt Sorenson
On Thu, Dec 12, 2002 at 05:17:29PM +, Dave Love wrote:
  Holt Sorenson [EMAIL PROTECTED] writes:
  
   I'm attempting to compile kernel 2.4.20 on a sun ultra 2 using 
   gcc-3.0 (3.0.4) from debian gnu/linux sparc woody 3.0r0.
  
  Don't.  Use `egcs64' -- see archives of this list.  Does the procedure

ah ha. The magic foo. Thanks! I tried searching the archives but
the search function on debian.org didn't return anything useful.

I tried things like 'kernel compile' and 'kernel'.

  in the kernel section of the Debian FAQ not work?  If it doesn't,
  please make a bug report.

Are you referring to 9.2 What tools does Debian provide to build custom
kernels? I haven't used it. I'm used to my old ways that grew from
using slackware. Thanks for pointing me towards a new trick. I'll give
it a go and adopt it if I'm happy with the process.

I just started using debian-sparc. I have used debian-i386 for quite
some time now, so I've had a few very minor bumps including this one
as I transition. Thanks for helping me knock one out.

  [What's the advantage of that kernel over the latest Debian package
  anyway?]

2.4.20 itself may or may not have significant advantages. I didn't
work really hard at comparing it to 2.4.18. I'm patching it
to support filesystem acls and exploring patching gr-security
into it. With patches that aren't part of the mainstream kernel,
it has been my experience that using the most recent patch with
the most recent kernel is generally a good idea.

   The linux kernel faq 
   (http://www.wsinf.edu.pl/doc/doc-linux-html/FAQ/Linux-FAQ-8.html)
   suggests that the linker is using a static libc for linking.

  I don't think that's a kernel FAQ, despite the name.

Ack. Thanks.



Re: 2.4.20 kernel link failure: undefined reference to `_mcount'

2002-12-12 Thread Holt Sorenson
On Thu, Dec 12, 2002 at 02:17:24PM -0600, Roy Bixler wrote:
  On Thu, Dec 12, 2002 at 05:17:29PM +, Dave Love wrote:
   Holt Sorenson [EMAIL PROTECTED] writes:
   
I'm attempting to compile kernel 2.4.20 on a sun ultra 2 using 
gcc-3.0 (3.0.4) from debian gnu/linux sparc woody 3.0r0.
   
   Don't.  Use `egcs64' -- see archives of this list.  Does the procedure
   in the kernel section of the Debian FAQ not work?  If it doesn't,
   please make a bug report.
  
  I have successfully used the latest gcc-3.2.1 builds to compile kernel
  2.4.20 on a Sunblade 100 (Ultra IIe processor).  At least, the kernel
  has been up for more than 7 days without a problem.

Ack. I will go with egcs64 for now, but thanks for the info.

  -- 
  Roy Bixler [EMAIL PROTECTED]
  The University of Chicago Press