Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Cobalt RaQ/Qube Systems -- Testing Needed
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 7, 8, 9, 10, 11  Next  
Reply to topic    Gentoo Forums Forum Index Gentoo on Alternative Architectures
View previous topic :: View next topic  
Author Message
marmalade
n00b
n00b


Joined: 03 Dec 2004
Posts: 2
Location: Munich - Germany

PostPosted: Tue Dec 14, 2004 4:42 pm    Post subject: intalling qube 1 - tulip pci problem Reply with quote

after some trouble, i managed to get gentoo running on a qube 1.
the problem is, that the tulip driver does not work and because of the missing
serial console debugging is not very easy. with an other network card and a
new kernel gentoo is now running on the blue box :-)

i have mini howto here to make a diskless boot for installing gentoo on the qube

Redhatter wrote:

Are you using patched kernel sources? IIRC last time I tried sticking 2.4.20
on my Qube 2 without patching the kernel, I got a corrupted serial console
with no network running.

yes i used a patched kernel, i emerged the mips-sources with
Code:
USE=cobalt PROFILE_ARCH="cobalt" emerge mips-sources

see below for errors while loading the tulip driver.

Redhatter wrote:

Also, what is the exact problem you're getting with building a kernel
using crossdev? Just saying you can't get it to work isn't exactly helpful.

yes you are right it was not very helpfull sorry.

Redhatter wrote:

Code:
# crossdev --arch=mipsel -k


i get
Code:
No valid binutils packages were found

with specifying the exact packages it woks:
Code:
crossdev --arch=mipsel --vgcc=3.3.3-r6 --vbinutils=2.14.90.0.8-r1 --vglibc=2.3.4.20040619-r2


Redhatter wrote:

As always, if you strike troubles... we're here to help :-)

Ok, the only problem is tulip driver:
when loading the tulip module i get:
Code:

/lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o: init_module: No such device
Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters.
      You may find more information in syslog or the output from dmesg
/lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o: insmod /lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o failed
/lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o: insmod tulip failed


dmesg output:
Code:

Linux Tulip driver version 0.9.15-pre12 (Aug 9, 2002)
PCI: Enabling device 00:07.0 (0045 -> 0047)
PCI: Unable to reserve I/O region #1:80@10100000 for device 00:07.0
PCI: Unable to reserve I/O region #1:80@10100000 for device 00:07.0


lspci -v
Code:

0000:00:00.0 Host bridge: Marvell Technology Group Ltd.: Unknown device 4146 (rev 01)
        Flags: bus master, medium devsel, latency 64
        Memory at <unassigned> (32-bit, non-prefetchable) [size=32M]
        Memory at 02000000 (32-bit, non-prefetchable) [size=16M]
        Memory at 1c000000 (32-bit, prefetchable) [size=32M]
        Memory at 1f000000 (32-bit, prefetchable) [size=16M]
        Memory at 14000000 (32-bit, non-prefetchable) [size=4K]
        I/O ports at 14000000 [disabled] [size=4K]
        Expansion ROM at 1f000000 [disabled] [size=16M]

0000:00:07.0 Ethernet controller: Digital Equipment Corporation DECchip 21041 [Tulip Pass 3] (rev 21)
        Flags: bus master, medium devsel, latency 64, IRQ 13
        I/O ports at 10100000 [size=128]
        Memory at 00800080 (32-bit, non-prefetchable) [size=128]
        Expansion ROM at 90000000 [disabled] [size=256K]

0000:00:09.0 ISA bridge: VIA Technologies, Inc. VT82C586/A/B PCI-to-ISA [Apollo VP] (rev 27)
        Flags: bus master, medium devsel, latency 0

0000:00:09.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06) (prog-if 8a [Master SecP PriP])
        Flags: bus master, fast Back2Back, medium devsel, latency 64
        I/O ports at cc00 [size=16]

0000:00:09.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 02) (prog-if 00 [UHCI])
        Subsystem: VIA Technologies, Inc. (Wrong ID) USB Controller
        Flags: medium devsel
        I/O ports at 0300 [disabled] [size=32]

0000:00:0a.0 Ethernet controller: 3Com Corporation 3c905 100BaseTX [Boomerang]
        Flags: bus master, medium devsel, latency 64, IRQ 9
        I/O ports at 108000 [size=64]
        Expansion ROM at <unassigned> [disabled] [size=64K]


thanks!
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Wed Dec 15, 2004 12:13 am    Post subject: Re: intalling qube 1 - tulip pci problem Reply with quote

marmalade wrote:
after some trouble, i managed to get gentoo running on a qube 1.
the problem is, that the tulip driver does not work and because of the missing
serial console debugging is not very easy. with an other network card and a
new kernel gentoo is now running on the blue box :-)

i have mini howto here to make a diskless boot for installing gentoo on the qube


Ohh nice... I shall have a look.

marmalade wrote:
Redhatter wrote:

Are you using patched kernel sources? IIRC last time I tried sticking 2.4.20
on my Qube 2 without patching the kernel, I got a corrupted serial console
with no network running.

yes i used a patched kernel, i emerged the mips-sources with
Code:
USE=cobalt PROFILE_ARCH="cobalt" emerge mips-sources

see below for errors while loading the tulip driver.


Well, I'm in the habit of downloading direct from the Linux/MIPS CVS repository, then patching the kernel with patches from here. What you've done should be correct.

marmalade wrote:
Redhatter wrote:

Also, what is the exact problem you're getting with building a kernel
using crossdev? Just saying you can't get it to work isn't exactly helpful.

yes you are right it was not very helpfull sorry.

Redhatter wrote:

Code:
# crossdev --arch=mipsel -k


i get
Code:
No valid binutils packages were found

with specifying the exact packages it woks:
Code:
crossdev --arch=mipsel --vgcc=3.3.3-r6 --vbinutils=2.14.90.0.8-r1 --vglibc=2.3.4.20040619-r2


Okay, does it work if you drop off the --vgcc, --vbinutils and --vglibc lines? If you really wish to use these, then you might have to tell it to use those versions reguardless of stability. Specify --unstable (-u) to do this.

marmalade wrote:

Ok, the only problem is tulip driver:
when loading the tulip module i get:
Code:

/lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o: init_module: No such device
Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters.
      You may find more information in syslog or the output from dmesg
/lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o: insmod /lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o failed
/lib/modules/2.4.27-mipscvs-20040814/kernel/drivers/net/tulip/tulip.o: insmod tulip failed


dmesg output:
Code:

Linux Tulip driver version 0.9.15-pre12 (Aug 9, 2002)
PCI: Enabling device 00:07.0 (0045 -> 0047)
PCI: Unable to reserve I/O region #1:80@10100000 for device 00:07.0
PCI: Unable to reserve I/O region #1:80@10100000 for device 00:07.0


lspci -v
Code:

0000:00:00.0 Host bridge: Marvell Technology Group Ltd.: Unknown device 4146 (rev 01)
        Flags: bus master, medium devsel, latency 64
        Memory at <unassigned> (32-bit, non-prefetchable) [size=32M]
        Memory at 02000000 (32-bit, non-prefetchable) [size=16M]
        Memory at 1c000000 (32-bit, prefetchable) [size=32M]
        Memory at 1f000000 (32-bit, prefetchable) [size=16M]
        Memory at 14000000 (32-bit, non-prefetchable) [size=4K]
        I/O ports at 14000000 [disabled] [size=4K]
        Expansion ROM at 1f000000 [disabled] [size=16M]

0000:00:07.0 Ethernet controller: Digital Equipment Corporation DECchip 21041 [Tulip Pass 3] (rev 21)
        Flags: bus master, medium devsel, latency 64, IRQ 13
        I/O ports at 10100000 [size=128]
        Memory at 00800080 (32-bit, non-prefetchable) [size=128]
        Expansion ROM at 90000000 [disabled] [size=256K]

0000:00:09.0 ISA bridge: VIA Technologies, Inc. VT82C586/A/B PCI-to-ISA [Apollo VP] (rev 27)
        Flags: bus master, medium devsel, latency 0

0000:00:09.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06) (prog-if 8a [Master SecP PriP])
        Flags: bus master, fast Back2Back, medium devsel, latency 64
        I/O ports at cc00 [size=16]

0000:00:09.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 02) (prog-if 00 [UHCI])
        Subsystem: VIA Technologies, Inc. (Wrong ID) USB Controller
        Flags: medium devsel
        I/O ports at 0300 [disabled] [size=32]

0000:00:0a.0 Ethernet controller: 3Com Corporation 3c905 100BaseTX [Boomerang]
        Flags: bus master, medium devsel, latency 64, IRQ 9
        I/O ports at 108000 [size=64]
        Expansion ROM at <unassigned> [disabled] [size=64K]


thanks!


That does look nasty. You might want to try moving up to Linux 2.6.9. IIRC Peter Horton managed to find the bug that was stopping this from working. As far as stability goes on 2.6, I've had no worries running 2.6.8.1 on my Qube 2, so 2.6.9 shouldn't be a problem.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
crahen
n00b
n00b


Joined: 11 Jan 2004
Posts: 51

PostPosted: Tue Dec 21, 2004 9:53 pm    Post subject: Reply with quote

This might be a silly queston, but what flags are masking mips-headers-2.6.* ebuilds? I've looked in the ebuilds themselves and it looks like just ~mips unless I'm completely missing something - which I seem to be because thats not working ....
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 367
Location: Sigma 957

PostPosted: Wed Dec 22, 2004 12:13 pm    Post subject: Reply with quote

crahen wrote:
This might be a silly queston, but what flags are masking mips-headers-2.6.* ebuilds? I've looked in the ebuilds themselves and it looks like just ~mips unless I'm completely missing something - which I seem to be because thats not working ....


mips-headers-2.6.* are masked because they are extremely experimental, and really only for testing of the n32 userland stuff when one of us gets spare time (mine has been extremely limited lately). As such, it is highly advised you steer clear of them. 2.6.10 may change some things, and maybe, make 2.6 somewhat viable for o32, but that's a completely uneducated guess, made in part on the vast number of improvements gone into linux-mips.org CVS since 2.6.9.

--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
crahen
n00b
n00b


Joined: 11 Jan 2004
Posts: 51

PostPosted: Wed Dec 22, 2004 1:34 pm    Post subject: Reply with quote

How was it masked? I'm wondering what it is I missed that I wasn't able to unmask it as I have other packages.

Also, has anyone ever tried to put an older, lower power VGA pci card into a Qube?
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 367
Location: Sigma 957

PostPosted: Wed Dec 22, 2004 7:42 pm    Post subject: Reply with quote

Portage has multiple layers from which you can control what packages are installed on a given system. Sometimes it can get complex. mips-headers-2.6, for example, is masked at the profile level. Because I didn't want normal o32 users to wind up with dead systems, I masked mips-headers-2.6* in all mips profiles except the n32 cascade profile under mips64. You've been unmasking packages at the keyword level mostly, which is what most users do, as profile-level masking is usually for tying a profile down to specific versions of software. All masking can be controlled by the user by setting up various files in /etc/portage (this includes profiles as well).

If you're interested in learning more about portage's control mechanisms, you'll want to peruse the docs on the main site more. as well as the man pages (man portage). There's also a guide on cascaded profiles as well.

Lastly, I still advise not using mips-headers-2.6, if that's what you're planning :)


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
crahen
n00b
n00b


Joined: 11 Jan 2004
Posts: 51

PostPosted: Thu Dec 23, 2004 3:55 am    Post subject: Reply with quote

I'll stick with the 2.4 headers for now. I'm still stuck getting libc and gcc to emereg ecorrectly - gas errors "operation spans multiple segments" but for the most part things seem to be working..
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 367
Location: Sigma 957

PostPosted: Thu Dec 23, 2004 7:20 am    Post subject: Reply with quote

That's a rather odd error. Can't say I've bumped into it before. What gcc, binutils, and glibc does the system currently have?


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
crahen
n00b
n00b


Joined: 11 Jan 2004
Posts: 51

PostPosted: Thu Dec 23, 2004 7:15 pm    Post subject: Reply with quote

Emerging glibc-2.3.4.20041102 using

binutils-2.15.91.0.2
glibc-2.3.3_pre20040420
gcc-3.3.3_20040412

Code:

{standard input}: Assembler messages:
{standard input}:131: Error: operation combines symbols in different segments
make[2]: *** [/var/tmp/portage/glibc-2.3.4.20041102/work/build-mipsel-unknown-li
nux-gnu-linuxthreads/signal/sigsuspend.o] Error 1
make[2]: Leaving directory `/var/tmp/portage/glibc-2.3.4.20041102/work/glibc-2.3
.3/signal'
make[1]: *** [signal/subdir_lib] Error 2
make[1]: Leaving directory `/var/tmp/portage/glibc-2.3.4.20041102/work/glibc-2.3
.3'
make: *** [all] Error 2


I got the same sort of error emerging ncurses.[/code]
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 367
Location: Sigma 957

PostPosted: Fri Dec 24, 2004 9:33 am    Post subject: Reply with quote

Few things:

One, that's a glibc error you pasted, not an ncurses. Two, you should be using gcc-3.4.x (gcc-3.4.1-r3 is stable on mips). Check with gcc-config -l.

Second, it looks like you're trying to use the unstable packages, which doesn't really carry much of a guarantee with it. If you continue to have problems, you might consider dropping to the stable package set, try and get that working, backup data if working, and then try and upgrade to unstable again.

Somethings to check:

1. Do you have your make.profile set to the correct profile? Current profile for cobalt hardware is /usr/portage/profiles/default-linux/mips/cobalt/2004.2

2. Check CFLAGS for anything insane. Insane things include any non-mips options (happens sometimes), -O3 (-O2 is the reccommended setting), and silly things like -falign-* stuff.

Most likely, you've probably got some packages merged incorrectly, or the wrong versions merged. Try some of the above ideas and see if anything works or not (you might consider starting from a stage1/2/3).


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
Astral
n00b
n00b


Joined: 17 Nov 2003
Posts: 19
Location: Mo, USA

PostPosted: Tue Dec 28, 2004 3:50 am    Post subject: Reply with quote

Here is a good one for you guys:

LD init/built-in.o
LD .tmp_vmlinux1
arch/mips/cobalt/built-in.o(.text+0x328): In function `cobalt_machine_halt':
: undefined reference to `led_state'
arch/mips/cobalt/built-in.o(.text+0x330): In function `cobalt_machine_halt':
: undefined reference to `led_state'
arch/mips/cobalt/built-in.o(.text+0x344): In function `cobalt_machine_halt':
: undefined reference to `led_state'
arch/mips/cobalt/built-in.o(.text+0x358): In function `cobalt_machine_halt':
: undefined reference to `led_state'
arch/mips/cobalt/built-in.o(.text+0x370): In function `cobalt_machine_halt':
: undefined reference to `led_state'
arch/mips/cobalt/built-in.o(.text+0x384): more undefined references to `led_stat
e' follow
make: *** [.tmp_vmlinux1] Error 1

I get that after compiling my 2.6.9 kernel. I usesd kumbas .config, and tweaked a few things. Any help will be appreciated.
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 367
Location: Sigma 957

PostPosted: Wed Dec 29, 2004 2:13 am    Post subject: Reply with quote

It sounds like you left something out of the .config. Below is a link to my current 2.6.9 kernel that's been running for 44 days now. Be aware that it's configured for udev and not devfs, so you will need to make some changes.

http://dev.gentoo.org/~kumba/mips/cobalt/kernel/cobalt-config-2609


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
Astral
n00b
n00b


Joined: 17 Nov 2003
Posts: 19
Location: Mo, USA

PostPosted: Thu Dec 30, 2004 1:52 am    Post subject: Reply with quote

Thank you, all wise all knowing Kumba, that worked great. I went ahead and emerged udev and just used that. Now running on my own kernel. I have a few packages I want to run. Let me know if it am crazy:

Apache2
Mysql
Tomcat (includes blackhawk java)
Postfix
IpTables (anything related to NAT, I want this to be my gateway)

Tomcat is masked, and never has been availalbe on mips, that I can tell. Am I dreaming or do you think I can do this? Thanks guys.
_________________
We need to start colonizing the moon.

-Astral-
Back to top
View user's profile Send private message
Sanguinary
n00b
n00b


Joined: 17 Oct 2004
Posts: 16

PostPosted: Thu Dec 30, 2004 3:03 am    Post subject: Reply with quote

Astral wrote:
I have a few packages I want to run. Let me know if it am crazy:

Apache2
Mysql
Tomcat (includes blackhawk java)
Postfix
IpTables (anything related to NAT, I want this to be my gateway)


Can't help with everything on your list, but I have built some of it, although I'm far from the latest ebuilds on some of these. (I'm on a 56K modem, so I emerge the ebuild for which I have already downloaded the source, unless I'm aware of some serious problem.)

Apache2 runs fine (running 2.0.51-r1).

Mysql (4.0.20-r1) built with no problem, but I decided to run Mysql on a different machine. When I was running NetBSD, Mysql was painfully slow.

Postfix runs fine (running 2.0.19).
Back to top
View user's profile Send private message
Kumba
Developer
Developer


Joined: 16 Jul 2002
Posts: 367
Location: Sigma 957

PostPosted: Thu Dec 30, 2004 7:14 am    Post subject: Reply with quote

Astral wrote:
Thank you, all wise all knowing Kumba, that worked great. I went ahead and emerged udev and just used that. Now running on my own kernel. I have a few packages I want to run. Let me know if it am crazy:

Apache2
Mysql
Tomcat (includes blackhawk java)
Postfix
IpTables (anything related to NAT, I want this to be my gateway)


These packages should work fine. MySQL is probably the only one that might be an issue, as it includes a copy of BerkeleyDB for some feature it needs, and this bdb copy can utilize a function due to a bug/missing feature on the mips arch (yeah, pretty vague, see the comments in the db/mySQL ebuilds, one of them has it). I forget what the possible consequences are, just slower operation I think. Been awhile.

Anyways, that aside, the rest, save Tomcat should be fine. I personally haven't tried Tomcat, and know very little about it. If it involves Java, you're gonna have fun, because mips doesn't have a working, full-blown KVM or jdk. I believe kaffe may be the only one that'd work, but big ones like blackdown, sun, ibm, etc.., won't run on mips.

psi29a, our resident KDE/webapps dev randomly pops out of the aether on IRC to update some of these packages, so if you fire him an email, you might get some more info about the current status of them.


--Kumba
_________________
"The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Back to top
View user's profile Send private message
rprescott
n00b
n00b


Joined: 09 Apr 2004
Posts: 16

PostPosted: Sun Jan 09, 2005 6:42 am    Post subject: Lost Reply with quote

Hi,

I got a hold of a RAQ4 machine and want to put Gentoo Linux on it but, I dont know where to being. I keep reading and finding more not clear or dont understand the information. Could you someone explain to me how to get Gentoo on a RAQ4i machine? I put the hard drive in another machine and installed gentoo....everything I think works expect I get agetty errors and cannot log into ssh or webmin and dont know if its the kernel?? Again lost?

Thanks,
Rob
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Sun Jan 09, 2005 9:00 am    Post subject: Re: Lost Reply with quote

rprescott wrote:
Hi,

I got a hold of a RAQ4 machine and want to put Gentoo Linux on it but, I dont know where to being. I keep reading and finding more not clear or dont understand the information. Could you someone explain to me how to get Gentoo on a RAQ4i machine? I put the hard drive in another machine and installed gentoo....everything I think works expect I get agetty errors and cannot log into ssh or webmin and dont know if its the kernel?? Again lost?


Hate to tell you this... but there's a big difference between a Raq/Qube 2, and a Raq 4. The Raq/Qube 2 runs a QED RM5231 RISC CPU running at 250MHz. The Raq 4 is a x86 machine running an AMD K6 CPU (either that or an Intel Pentium III). So if you're following the tips here, undoubtedly you'll be confused.

I have some bits & pieces on the x86-based Qube/Raq machines here. Some others have managed to get Gentoo going on Qube 3's on this forum, so you might want to look around and do a forum search.

AFAIK you need custom kernel patches to make the networking interface work correctly. Beyond that, I wouldn't have a clue as I don't use one of these machines myself.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
erikdenv
n00b
n00b


Joined: 24 May 2004
Posts: 30
Location: The Netherlands

PostPosted: Mon Jan 17, 2005 12:13 pm    Post subject: Reply with quote

Hello, I recently got a Qube2 from a friend and thanks to this thread and your help I was able to get Gentoo running on it.

The Qube runs perfect but I can only seem to get it booted with net booting and only when it's attached to my server.
When I monitor the booting process I see the Qube is trying to get it's address from my server with DHCP, it's trying to get it's previous address (192.168.0.16) from my server (192.168.0.100). (this is even before loading the Linux kernel)

When I attach it to the ADSL-modem the system won't boot at all.
My DHCP-server supplies addresses based on MAC-address, when I change the address of the Qube from x.x.x.16 to x.x.x.17 the Qube also won't boot anymore.

What am I doing wrong?
I want to replace the current server with the Qube so I would be nice if it could boot by itself :-)
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Mon Jan 17, 2005 2:05 pm    Post subject: Reply with quote

erikdenv: Okay, By "The Qube runs perfect but I can only seem to get it booted with net booting..." do you mean that the machine is not booting from its internal hard drive?

If this is the case, then chances are your ADSL router isn't giving the machine the NFS root path & other bits and pieces it needs to boot.

The Qube shouldn't even need to be given an IP number prior to booting the kernel under normal circumstances. Are you using the CoLo firmware or stock firmware? Also, are any buttons being pressed when the machine is switched on?

If you can get a dump of the bootup messages whilst the machine is being powered on -- this might help in understanding the problem.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
erikdenv
n00b
n00b


Joined: 24 May 2004
Posts: 30
Location: The Netherlands

PostPosted: Mon Jan 17, 2005 8:52 pm    Post subject: Reply with quote

Hello, I will rephrase my problem, hope is better readable this way

This is the working variant:

The Qube is connected to my server
I use the standard bootloader
I installed colo and renamed one of the files to vmlinux as written in the manual
I can monitor the boot process with xc over the serial cable
If I boot the cube and press the left and right button the Qube starts netbooting
The bootloader starts the colo.default script and loads the kernel from the harddisk
The Qube boots fine

It seems the Qube does a DHCP request even before the kernel has started, I guess it's because I use net booting

This doesn't work:
1. I just boot the Qube without pressing any buttons, the display says "starting up and that's it" can't see anything with xc (it should be possible to just boot it, shouldn't it?)
2. I connect the Qube to the ADSL modem and use net booting, can't see anything with x
3. I change the address for the Qube in the dhcp-server and try netbooting, can't see anything with x

I'll make bootlogs as soon as I get to it again
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Tue Jan 18, 2005 1:18 am    Post subject: Reply with quote

Ahh... did you format /dev/hda1 as EXT2 revision 0?

The stock firmware, when it boots, expects to see an EXT2r0 partition on /dev/hda1, containing a file called 'vmlinux.gz' that is less than 675kB.

Try holding in the "password reset" button whilst switching the unit on -- Console ON should appear on the back LCD, then just power cycle it again -- you'll see the cobalt networks diagnostics come up before it finally mounts the partition and tries to kickstart the kernel (in your case, CoLo).

What's your partition layout look like? Do you have a separate partition for /boot or do you just put it in /?

Here are some hints you might like to look at for getting your box going.
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
erikdenv
n00b
n00b


Joined: 24 May 2004
Posts: 30
Location: The Netherlands

PostPosted: Tue Jan 18, 2005 10:26 am    Post subject: Reply with quote

Aha, yes I have a separate boot partition and yes I did format it with Ext2 but not specificly with revision 0. I had read that this was not necessary with Colo but that's only true when you flash the old bootloader.

I guess it's best when I reformat my /boot partion with ext2 revision 0

thank you for your help

edit: It's working now, it boots perfectly, thanks


Last edited by erikdenv on Wed Jan 19, 2005 6:36 am; edited 2 times in total
Back to top
View user's profile Send private message
atopian
n00b
n00b


Joined: 18 Apr 2002
Posts: 29
Location: Columbus, OH

PostPosted: Wed Jan 19, 2005 3:13 am    Post subject: Reply with quote

The thing about the crashing with some mips cobalt boxen is a hardware IRQ conflict between the serial interface and the ethernet port. The easiest fix is to stick a daemon on the serial interface that never actually sends or recieves any data. I personally use pppd and disable the serial console. Yea, its kinda crappy, but if you have a NFS environment around one can work around the drawback.

For some reason these issues also manifests when running a hardware level dhcpc.

This is also the reason btw, why the debian team allows you to telnet into the boxes during the install. Otherwise crashy crashy. Ive had friends keep their mips cobalt boxen up for months by doign the pppd trick. LIke I said though the key is making sure ppp dosent try to send any data (keepalive etc).
_________________
Gentoo, boldly going places only NetBSD dare venture.
Back to top
View user's profile Send private message
Redhatter
Retired Dev
Retired Dev


Joined: 20 Sep 2003
Posts: 548
Location: Brisbane, QLD, Australia

PostPosted: Wed Jan 19, 2005 4:32 am    Post subject: Reply with quote

atopian: Depends on your kernel though. Linux 2.6.8.1 has been quite stable on my file server, with both serial and network activity:

Code:
(14:30) stuartl@beast stuartl $ ssh qube uptime
 16:40:33 up 4 days,  6:40,  0 users,  load average: 0.00, 0.00, 0.00


... I have had that box running non stop for as long as a fortnight. The only thing that crashes it at the moment, is one dickey SCSI HDD (I have an Adaptec SCSI card installed with 4 external drives).

What kernel are you using? Perhaps its worth an update... ;-)
_________________
Stuart Longland (a.k.a Redhatter, VK4MSL)
I haven't lost my mind - it's backed up on a tape somewhere...

Gentoo/MIPS Cobalt developer, Mozilla herd member.
Back to top
View user's profile Send private message
Francis85
n00b
n00b


Joined: 27 Jan 2003
Posts: 35

PostPosted: Mon Jan 24, 2005 3:33 am    Post subject: Reply with quote

I updated to 2.6.10, I did not omit the Cobalt LCD driver, but when I use putlcd I get this

Writing LCD: No such file or directory

Any way to fix it?
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo on Alternative Architectures All times are GMT
Goto page Previous  1, 2, 3 ... 7, 8, 9, 10, 11  Next
Page 8 of 11

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum