linux-apus-user

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

> Just Ctrl-Alt-F2 and there should be a shell IIRC.

I will try this. Although I think my biggest problem whilst configuring the
base system was knowing what I wanted and what I didn't!! Not sure how to
figure that out yet. In theory I would like to leave as much installed as
possible and know what to install to get some sort of gui and maybe some
apps like a browser etc installed - but I may be trying to run before I can
crawl here...... not forgetting that the clock problem isn't really resolved
yet (I'm reminding myself). 

I remember getting some dependency conflicts and not being able to resolve
them. Maybe I will post details here later after I have had another go. 

> This is good news! Also that this clearly isolates the system crashes to
> the hardware clock reads.

Yes, I am glad that the problem seems to have been totally isolated. Thanks
in no small part to you - nobody had made those suggestions before and this
is the furthest I have ever got with trying to get Linux running on my
machine. 

>> > But this is definitely not how a productive system should be run! So,
>> > try to get the dmesg output anyway and post it.

I will put the clock stuff back to how it was originally and collect some
debug output to post here. Watch this space.  

> Because you didn't tell it to save the stuff in the memory. Well, I
> guess it was me who forgot to mention it in the last mail anyway..

Don't worry, better late than never. And if my brain had retained any of the
stuff I had been reading, I would have noticed that myself.  

>> The other question is, is there a way
>> of resetting the Linux side which cleanly unmounts my hda2 partition?

> After the crash? I don't think so. Except if you can get a debug monitor
> running maybe, but this is over my head ATM.

No, I meant before a crash. For example when running the installation
program, if I chose "reboot the system" then when starting Linux again it
would run a check on hda2 everytime because it would complain that the
partition had not been cleanly unmounted. I am sure I even "unmounted" hda2
and still the revalidate ran afterwards - I will recheck that but I'm sure
that's what happened.

Thank you for all your help so far - I really appreciate it. 

Kind regards, 

Kulwant

Please check your mailing list prefs

[John L. <ch...@la...>]

Hello Linux-APUS-user,

  I unsubbed quite a while ago from this list, but although not getting
  'list' messages, am still getting some from some members.

  Please everyone, delete me from whatever settings you have for this list.
  My miggy died and I no longer own one.

-- 
Best regards,
 John                          mailto:ch...@la...

Re: Version(s) working with BlizzardPPC

[Geert U. <ge...@li...>]

On Sun, 14 Dec 2003, Andreas [ISO-8859-1] W=FCst wrote:
> On Sun, 2003-12-14 at 04:20, Kulwant Bhogal wrote:
> > > If it doesn't work yet, you could also rename the hwclockfirst.sh c=
all,
> > > although it doesn't seem to write to the clock.
> >
> > >      # cd etc/rcS.6/
> > >      # mv S18hwclockfirst.sh offS18hwclockfirst.sh
> >
> > > But I would heavily advise against it, because then nothing is sett=
ing
> > > the system clock (system clock !=3D hardware clock), which could le=
ad to
> > > strange behaviour of the system.
> >
> > Well, I did try that (because all else before that had failed) and I =
got:
> >
> > Congratulations, You have successfully installed
>
> Yeah, so we've got one step further! :)
>
> > and after the message about how to revisit the setup process (by runn=
ing
> > /usr/sbin/base-config) I got another kernel panic:
> >
> > Loading /etc/console/boottime.kmap.gz
> >
> > kernel access of bad area pc c00dca04 lr c00de9f8 address 70 tsk hwcl=
ock/230
                                                              ^^
Looks like a NULL-pointer struct dereference.

Do you have a System.map? Can you look up the label (closest before) for
c00dca04 and c00de9f8?

> > task =3D c65dc000[230] 'hwclock'
>
> So you really seem to have a major problem with the hw clock!
>
> Is the contact between the ppc board and the motherboard ok? Where does
> the clock actually reside (motherboard or ppc)? Is it some strange
> non-standard chip?

How does the kernel detect your clock? A2000 or A3000 style?

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m6=
8k.org

In personal conversations with technical people, I call myself a hacker. =
But
when I'm talking to journalists I just say "programmer" or something like=
 that.
							    -- Linus Torvalds

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Sun, 2003-12-14 at 23:16, Kulwant Bhogal wrote:

> Hello Andi, 
> 
> > Though in /drivers/char/rtc.c there's a comment mentioning something
> > about periodic syncs (every 11 min) by time.c (line 54).. So you're
> > system should at least run for 11 min uninterrupted..
> 
> See below. 
> 
> > What you could also do is to
> 
> >   # mv /sbin/hwclock /sbin/OFFhwclock
> 
> > although this is really really ugly!! But it should prevent the system
> > from interacting with the hardware clock.
> 
> Ok, I did this. And no crashes.

Cool!

> It went on to configuring the base system
> etc. But I didn't really know what I was doing so I have quit out of that
> without installing or deleting anything. I didn't get to a shell either so I
> couldn't set the date even manually (as you suggested below).

Just Ctrl-Alt-F2 and there should be a shell IIRC.

> > Afterwards set the system clock using
> 
> >   # date 1214200003
> 
> Well the system did run in the configure base system section for ages (I
> left it alone for over an hour) and I didn't get any problems associated
> with the 11 min interupt you suggested above.

This is good news! Also that this clearly isolates the system crashes to
the hardware clock reads.

> > But this is definitely not how a productive system should be run! So,
> > try to get the dmesg output anyway and post it.
> 
> Try as I might I just don't get any debug output from dmesg. It simply says:
> 
> Searching for SAVEBOOT magic...
> Not found
> 
> I have tried starting wth no startup-sequence and before the reset from
> OS3.9 but still no joy. I inserted the -d in my start script like so: 
> 
> bootstrap --apus -k linux.bin -d -v root=/dev/hda2 rw nobats video=pm2fb:mode:1024x768-75
> 
> not sure what the -v is for.

The -v option stands for verbose IIRC (I guess it activates some
additional information for boothack, NOT the kernel itself).

> Is the -d in the right place?

Don't know about the place, but there's definitely something lacking,
you forgot to add the debug=mem part (or better, I didn't tell you about
it..). Please do like this:

bootstrap --apus -k linux.bin -v root=/dev/hda2 rw nobats
video=pm2fb:mode:1024x768-75 -d debug=mem

C.f. http://linux-apus.sourceforge.net/FAQ.html for full description.

> I did notice that I do get more prompts on running that script with the -d
> option in there.

I guess that's correct.

> So I think it is doing something. But Dmesg doesn't find
> any debug information after a reset.

Because you didn't tell it to save the stuff in the memory. Well, I
guess it was me who forgot to mention it in the last mail anyway..

> The other question is, is there a way
> of resetting the Linux side which cleanly unmounts my hda2 partition?

After the crash? I don't think so. Except if you can get a debug monitor
running maybe, but this is over my head ATM.

> And is there a way of resetting which will save the debug information too?

Well, I think debug=mem will do. It was just that this was lacking.
Sorry for that!

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

> Though in /drivers/char/rtc.c there's a comment mentioning something
> about periodic syncs (every 11 min) by time.c (line 54).. So you're
> system should at least run for 11 min uninterrupted..

See below. 

> What you could also do is to

>   # mv /sbin/hwclock /sbin/OFFhwclock

> although this is really really ugly!! But it should prevent the system
> from interacting with the hardware clock.

Ok, I did this. And no crashes. It went on to configuring the base system
etc. But I didn't really know what I was doing so I have quit out of that
without installing or deleting anything. I didn't get to a shell either so I
couldn't set the date even manually (as you suggested below). 

> Afterwards set the system clock using

>   # date 1214200003

Well the system did run in the configure base system section for ages (I
left it alone for over an hour) and I didn't get any problems associated
with the 11 min interupt you suggested above. 

> But this is definitely not how a productive system should be run! So,
> try to get the dmesg output anyway and post it.

Try as I might I just don't get any debug output from dmesg. It simply says:

Searching for SAVEBOOT magic...
Not found

I have tried starting wth no startup-sequence and before the reset from
OS3.9 but still no joy. I inserted the -d in my start script like so: 

bootstrap --apus -k linux.bin -d -v root=/dev/hda2 rw nobats video=pm2fb:mode:1024x768-75

not sure what the -v is for. Is the -d in the right place?  

I did notice that I do get more prompts on running that script with the -d
option in there. So I think it is doing something. But Dmesg doesn't find
any debug information after a reset. The other question is, is there a way
of resetting the Linux side which cleanly unmounts my hda2 partition? 
And is there a way of resetting which will save the debug information too? 

Kind Regards, 

Kulwant

Re: Debian refuses to boot!

[Daniel I. R. <ir...@ip...>]

Hello a.l...@lb...

> that amikeyb line is just the ctrl-amiga-amiga keycombo - its the few
> lines above that are interesting... last_ipl - some interupts arent being
> serviced properly.

> are you running the same kernel as before (i dont have your original mail)
> ?? you've upgraded your debian, havent you? 

No, I haven't... 

     Well, my installation is working now...  I'm using the latest kernel, 
(2.4.20). Just have a few questions:

     Where should I icopy the module files? Is it /lib/modules/2.4.20-apus
or /lib/modules/2.4.20?

     Where should I put the system.map file?

     FInally, I installed Xfree86 and KDE with apt-install kde..and copied
the file XF86Config-4 in /etc/X11 (which I downloaded from the apus site)
but when I run the command startx , XWindow does not start..  (see attached
logfile). I tried with the other file in the apus site: XF86Config-4-2 and
X didn't start either however the log file seems to report another problem
(attaching file too) . The logs correspond to XF86Config-4 and
XF86Config-4-2 respectively. Is there something I'm missing? 

     Any help will be appreciated!

Thanks!!

Daniel

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

> Which means it already crashes when trying to simply READ the clock.

Yes, it would appear so. 

>> Loading /etc/console/boottime.kmap.gz
>> 
>> kernel access of bad area pc c00dca04 lr c00de9f8 address 70 tsk
>> hwclock/230 task = c65dc000[230] 'hwclock'

> /sbin/hwclock tries to read the time from the RTC. You have to know that
> the system clock gets set by the hardware clock during boot, and
> afterwards the hardware clock is not going to be probed again (except
> during shutdown, when the hwclock.sh script tries to write back the
> system time to the hardware clock). The system clock generates it's own
> interrupts, so the hardware clock is mainly there to keep the time when
> the system is powered off. C.f. man hwclock.

Makes sense. 

> Though in /drivers/char/rtc.c there's a comment mentioning something
> about periodic syncs (every 11 min) by time.c (line 54).. So you're
> system should at least run for 11 min uninterrupted..

hmmm....

> I have no clue here. According to
> http://lxr.linux.no/diff/drivers/char/rtc.c?a=ppc;diffval=2.4.18;diffvar=v
> the rtc driver got changed between 2.4.18 and 2.4.19, but no difference
> between 2.4.19 and 2.4.20. Though there may also be some apus specific
> changes not documented here. The changes seem to have something to do
> with sysctl, though I don't know what exactly (sysctl stuff seems to got
> added to 2.4.19). 
> What you could also do is to

>   # mv /sbin/hwclock /sbin/OFFhwclock

> although this is really really ugly!! But it should prevent the system
> from interacting with the hardware clock.

> Afterwards set the system clock using

>   # date 1214200003

> (for December 14 (1214), 8 p.m. (2000), 2003 (03)). You'd have to do run
> date every time you boot your system.

> But this is definitely not how a productive system should be run! So,
> try to get the dmesg output anyway and post it.

I will try both and see what happens. 

Kind Regards, 

Kulwant

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

>> kernel access of bad area pc c00dca04 lr c00de9f8 address 70 tsk
>> hwclock/230 task = c65dc000[230] 'hwclock'
>                        ^^^^^^^^^

> So you really seem to have a major problem with the hw clock!

> Is the contact between the ppc board and the motherboard ok? Where does
> the clock actually reside (motherboard or ppc)? Is it some strange
> non-standard chip?

I believe the contact between the PPC board and motherboard is good,
otherwise I would expect instability on the Amiga side too. But the Amiga
side is as stable as rock. I can have the machine running all day with
IBrowse, SimpleMail, Amplifier (running WOS ppc mp3 decoding), MiamiDX,
Samba (for occasional networking to PCs), Jabberwocky, Editpad, Classaction,
Warp jpg Datatypes etc. in constant use with no lock-ups, crashes, freezes etc.     

I can even play Quake WOS and GL_Quake, Shogo, WipeOut and Freespace so if
there was a PPC problem, I would expect it to manifest itself here. Although Freespace
can occasionally lock-up, Shogo and Quake don't and neither does WipeOut, so I on that
front I suspect the software rather than the hardware. 

The clock resides on the PPC board. The board AFAIK is a standard Phase5
BlizzardPPC (second hand - but I've had it since the year 2000 and not had
any problems with it). The battery had nearly started to leak so I
unsoldered it and soldered in a CMOS battery holder and replaced the
battery. The clock still functions (i.e. saves and holds time) from the
AmigaOS side, so if I had blown anything I would expect it not to work from
the AmigaOS side either. I can open the Tower up and report the Revision 
number of the board (If I can find it) if it will help. 

> Can you add -d debug to your bootstrap line, and after the kernel oops
> reboot run dmesg > ram:output in the shell (amigaOS side)? Post the
> output here. I hope some of the kernel gurus can do something about it.

Off to try it now. 

> Well, the goal would be to run, not just crash every other minute..

But of course :-). 

Thanks for all your help so far. 

Kind Regards, 

Kulwant

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Sun, 2003-12-14 at 04:20, Kulwant Bhogal wrote:

> > If it doesn't work yet, you could also rename the hwclockfirst.sh call,
> > although it doesn't seem to write to the clock.
> 
> >      # cd etc/rcS.6/
> >      # mv S18hwclockfirst.sh offS18hwclockfirst.sh
> 
> > But I would heavily advise against it, because then nothing is setting
> > the system clock (system clock != hardware clock), which could lead to
> > strange behaviour of the system.
> 
> Well, I did try that (because all else before that had failed) and I got:

Which means it already crashes when trying to simply READ the clock.

> Loading /etc/console/boottime.kmap.gz
> 
> kernel access of bad area pc c00dca04 lr c00de9f8 address 70 tsk hwclock/230
> task = c65dc000[230] 'hwclock'

/sbin/hwclock tries to read the time from the RTC. You have to know that
the system clock gets set by the hardware clock during boot, and
afterwards the hardware clock is not going to be probed again (except
during shutdown, when the hwclock.sh script tries to write back the
system time to the hardware clock). The system clock generates it's own
interrupts, so the hardware clock is mainly there to keep the time when
the system is powered off. C.f. man hwclock.

Though in /drivers/char/rtc.c there's a comment mentioning something
about periodic syncs (every 11 min) by time.c (line 54).. So you're
system should at least run for 11 min uninterrupted..

> mm-> pgd c6e2c000 last syscall: 6
> 
> GPR00:   00001072  C65DDE10   C65DC000   00000001   00008000  00000000
> C6E09D80    C980000
> 
> GPR08:  0000000B   00000000   C0180000  C67EC08C   2422
> 
> and then it rebooted before I could write down the rest of it.

I have no clue here. According to
http://lxr.linux.no/diff/drivers/char/rtc.c?a=ppc;diffval=2.4.18;diffvar=v the rtc driver got changed between 2.4.18 and 2.4.19, but no difference between 2.4.19 and 2.4.20. Though there may also be some apus specific changes not documented here. The changes seem to have something to do with sysctl, though I don't know what exactly (sysctl stuff seems to got added to 2.4.19).

What you could also do is to

  # mv /sbin/hwclock /sbin/OFFhwclock

although this is really really ugly!! But it should prevent the system
from interacting with the hardware clock.

Afterwards set the system clock using

  # date 1214200003

(for December 14 (1214), 8 p.m. (2000), 2003 (03)). You'd have to do run
date every time you boot your system.

But this is definitely not how a productive system should be run! So,
try to get the dmesg output anyway and post it.

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Sun, 2003-12-14 at 04:20, Kulwant Bhogal wrote:

> > If it doesn't work yet, you could also rename the hwclockfirst.sh call,
> > although it doesn't seem to write to the clock.
> 
> >      # cd etc/rcS.6/
> >      # mv S18hwclockfirst.sh offS18hwclockfirst.sh
> 
> > But I would heavily advise against it, because then nothing is setting
> > the system clock (system clock != hardware clock), which could lead to
> > strange behaviour of the system.
> 
> Well, I did try that (because all else before that had failed) and I got: 
> 
> Congratulations, You have successfully installed 

Yeah, so we've got one step further! :)

> and after the message about how to revisit the setup process (by running
> /usr/sbin/base-config) I got another kernel panic: 
> 
> Loading /etc/console/boottime.kmap.gz
> 
> kernel access of bad area pc c00dca04 lr c00de9f8 address 70 tsk hwclock/230
> task = c65dc000[230] 'hwclock'
                       ^^^^^^^^^

So you really seem to have a major problem with the hw clock!

Is the contact between the ppc board and the motherboard ok? Where does
the clock actually reside (motherboard or ppc)? Is it some strange
non-standard chip?

> mm-> pgd c6e2c000 last syscall: 6
> 
> GPR00:   00001072  C65DDE10   C65DC000   00000001   00008000  00000000
> C6E09D80    C980000
> 
> GPR08:  0000000B   00000000   C0180000  C67EC08C   2422
> 
> and then it rebooted before I could write down the rest of it.

Can you add -d debug to your bootstrap line, and after the kernel oops
reboot run dmesg > ram:output in the shell (amigaOS side)? Post the
output here. I hope some of the kernel gurus can do something about it.

> So I guess you were right about it leading to strange behaviour of the
> system :-)

Well, this wasn't particularly what I've been thinking about w.r.t.
strange behaviour, because it seemed to crash again during an attempt to
read the hwclock.

> But it appears to be the furthest I have got so far :-)

Well, the goal would be to run, not just crash every other minute..

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Geert U. <ge...@li...>]

On Sun, 14 Dec 2003, Kulwant Bhogal wrote:
> Incidentally, I also tried out the 2.4.18 kernel and that freezes when it
> goes through the idebus section. I did notice that at the top it said 60MHz
> bus speed, but in the ide section it said 50MHz and mentioned overriding it
> with a "idebus=xx" option, I guess I do this in the bootstrap command as an
> option?

The idebus value shouldn't matter for us.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

> If it doesn't work yet, you could also rename the hwclockfirst.sh call,
> although it doesn't seem to write to the clock.

>      # cd etc/rcS.6/
>      # mv S18hwclockfirst.sh offS18hwclockfirst.sh

> But I would heavily advise against it, because then nothing is setting
> the system clock (system clock != hardware clock), which could lead to
> strange behaviour of the system.

Well, I did try that (because all else before that had failed) and I got: 

Congratulations, You have successfully installed Debian!

and after the message about how to revisit the setup process (by running
/usr/sbin/base-config) I got another kernel panic: 

Loading /etc/console/boottime.kmap.gz

kernel access of bad area pc c00dca04 lr c00de9f8 address 70 tsk hwclock/230
task = c65dc000[230] 'hwclock'
mm-> pgd c6e2c000 last syscall: 6

GPR00:   00001072  C65DDE10   C65DC000   00000001   00008000  00000000
C6E09D80    C980000

GPR08:  0000000B   00000000   C0180000  C67EC08C   2422

and then it rebooted before I could write down the rest of it. 

So I guess you were right about it leading to strange behaviour of the
system :-)

But it appears to be the furthest I have got so far :-) 

Kind Regards, 

Kulwant

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

> Hmm, since normal booting doesn't work for you, boot with the install
> image:

>  1) go directly to the step of the installer where you can mount
> partitions. There, you choose your root partition. IIRC it gets mounted
> under /target/.

>  2) go to the step where you can open a shell. Then
>      # cd /target/
>      # cd etc/rcS.d/
>      # mv S50hwclock.sh offS50hwclock.sh

>     "S50hwclock.sh" is in fact a link to "../init.d/hwclock.sh".

>      To prevent troubles during shutdown, do something similar for
> rc6.d:

>      # cd ../rc6.d/
>      # mv K25hwclock.sh offK25hwclock.sh

> BTW: I have no idea why it tries to write to the clock at boot time at
> all.

OK, tried the above and I still get a kernel panic (which I can't capture
with dmesg - it always says: 

"Searching for SAVEKMSG magic...
Not found"

I managed to write down this before it rebooted:

System time was Sun Dec 14 02:45:22 UTC 2003
Setting the System Clock using the Hardware Clock as reference...
NIP: C00DCA04 XER: 20000000 LR: C00DC988 REGS: c7c71d60 TRAP: 0300


There was a fair bit after that which I didn't catch. 

> If it doesn't work yet, you could also rename the hwclockfirst.sh call,
> although it doesn't seem to write to the clock.

>      # cd etc/rcS.6/
>      # mv S18hwclockfirst.sh offS18hwclockfirst.sh

> But I would heavily advise against it, because then nothing is setting
> the system clock (system clock != hardware clock), which could lead to
> strange behaviour of the system.

Hmmm, about to try this now.......

Kind Regards, 

Kulwant

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Sat, 2003-12-13 at 17:08, Kulwant Bhogal wrote:

> Yes, but I haven't been through all of the old kernels yet. I'm hoping to
> give that a go this weekend. Will report my findings. 
> 
> > Btw., which kernel did you use to install the system?
> 
> 2.4.20.

Interesting. Since it fails during writing to the harware clock, this
means you never got past the first phase of the installation?

> > Maybe you want to turn off the /etc/init.d/hwclock.sh script to
> > temporarily stop the system from writing to the clock registers during
> > startup/shutdown. It gets called in rcS.d as well as in rc6.d if I'm
> > right.
> 
> Haven't got a clue how to do that - not even where to start. Could you post
> step by step instructions please?

Hmm, since normal booting doesn't work for you, boot with the install
image:

 1) go directly to the step of the installer where you can mount
partitions. There, you choose your root partition. IIRC it gets mounted
under /target/.

 2) go to the step where you can open a shell. Then
     # cd /target/
     # cd etc/rcS.d/
     # mv S50hwclock.sh offS50hwclock.sh

    "S50hwclock.sh" is in fact a link to "../init.d/hwclock.sh".

     To prevent troubles during shutdown, do something similar for
rc6.d:

     # cd ../rc6.d/
     # mv K25hwclock.sh offK25hwclock.sh

BTW: I have no idea why it tries to write to the clock at boot time at
all.

If it doesn't work yet, you could also rename the hwclockfirst.sh call,
although it doesn't seem to write to the clock.

     # cd etc/rcS.6/
     # mv S18hwclockfirst.sh offS18hwclockfirst.sh

But I would heavily advise against it, because then nothing is setting
the system clock (system clock != hardware clock), which could lead to
strange behaviour of the system.

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andi, 

> Hmm, if I understand you right, 2.4.20 fails because of the clock stuff
> but pcmcia would work, and other kernels fail because of the apne stuff?

Yes, but I haven't been through all of the old kernels yet. I'm hoping to
give that a go this weekend. Will report my findings. 

> Btw., which kernel did you use to install the system?

2.4.20. 

> Maybe you want to turn off the /etc/init.d/hwclock.sh script to
> temporarily stop the system from writing to the clock registers during
> startup/shutdown. It gets called in rcS.d as well as in rc6.d if I'm
> right.

Haven't got a clue how to do that - not even where to start. Could you post
step by step instructions please? 

Thanks for all your advice so far. 

Kind regards, 

Kulwant

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Wed, 2003-12-10 at 23:16, Kulwant Bhogal wrote:

> Hello Andreas, 
> 
> Thanks for you mail.

You're welcome!

> > Use the latest boothack available.
> 
> yes that has helped in this instance.

Ok, so first problem solved. :)

> > Hmm, hasn't there been a problem in recent kernels with apne? I think it
> > got fixed in kernel 2.4.21 (search the apus devel archive for thread
> > "Kernel 2.4.21". The posts are from july/august 2003.).
> 
> err, where is the devel arch?

http://sourceforge.net/mailarchive/forum.php?forum_id=7161

And there finally seems to be a possibility to search the archive. :)

> > If it doesn't work at all, remove the pcmcia network card. Though I
> > guess you want some kind of ethernet connection.. ;)
> 
> Well the PCMCIA ethernet problem doesn't surface in 2.4.20 so I will try
> Alan's suggestion of an older kernel and if that solves the clock problem
> then I will have to tackle whatever other problems that throws up as and
> when they occur.

Hmm, if I understand you right, 2.4.20 fails because of the clock stuff
but pcmcia would work, and other kernels fail because of the apne stuff?

Btw., which kernel did you use to install the system?

Maybe you want to turn off the /etc/init.d/hwclock.sh script to
temporarily stop the system from writing to the clock registers during
startup/shutdown. It gets called in rcS.d as well as in rc6.d if I'm
right.

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Wed, 2003-12-10 at 23:48, Kulwant Bhogal wrote:

> Hello Alan, 
> 
> Thanks for your mail. 
> 
> >> Is someone working on 2.4.23? I do find it strange that an older version
> >> is more compatible than a new version. I also wonder how many versions
> >> of the
> 
> > many many changes were being undertaken to the kernel during the 2.4.x
> > lifetime and each change usually killed something that wasnt/isnt being
> > maintained (look at the serial interface support etc) 
> 
> That sounds a bit sad. Why is that, surely not to limit the size to fit on a
> 1.44Mb floppy or something as daft as that?

No. Normal updates to drivers etc. Mostly adding support for newer
machines (ok, besides vm changes.. ;) ).

> I can't understand why adding
> new features has to disturb existing/working features - especially if they
> have no common bits (I mean what has settting the clock while booting got to
> do with serial i/f support????).

I think he was just talking about serial support as an example for
something being broken during the evolution of the 2.4.x kernel series.

Problem is, that code which is shared by many different
machines/architectures (think general stuff) gets altered e.g. to
accommodate new hardware. This mostly works for the commonly used
architectures since these people work with x86 etc. and there are a lot
of people testing this stuff. Unfortunately, these modifications every
now and then break the not so common architectures (be it because they
use preprocessor switches to in/exclude particular lines of code, and
x86 developers don't know about the impact their changes have on other
architectures, etc.). Since we are not tracking the rc kernels, no one
can test this stuff and bug the developer in charge for the change, and
if something gets broken, there's also noone to fix it..

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Alan, 

Thanks for your mail. 

>> Is someone working on 2.4.23? I do find it strange that an older version
>> is more compatible than a new version. I also wonder how many versions
>> of the

> many many changes were being undertaken to the kernel during the 2.4.x
> lifetime and each change usually killed something that wasnt/isnt being
> maintained (look at the serial interface support etc) 

That sounds a bit sad. Why is that, surely not to limit the size to fit on a
1.44Mb floppy or something as daft as that? I can't understand why adding
new features has to disturb existing/working features - especially if they
have no common bits (I mean what has settting the clock while booting got to
do with serial i/f support????). 

Rgds

Kulwant

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Andreas, 

Thanks for you mail. 

> Use the latest boothack available.

yes that has helped in this instance. 

> Hmm, hasn't there been a problem in recent kernels with apne? I think it
> got fixed in kernel 2.4.21 (search the apus devel archive for thread
> "Kernel 2.4.21". The posts are from july/august 2003.).

err, where is the devel arch? 

> If it doesn't work at all, remove the pcmcia network card. Though I
> guess you want some kind of ethernet connection.. ;)

Well the PCMCIA ethernet problem doesn't surface in 2.4.20 so I will try
Alan's suggestion of an older kernel and if that solves the clock problem
then I will have to tackle whatever other problems that throws up as and
when they occur. 

Kind regards, 

Kulwant

Re: Version(s) working with BlizzardPPC

[Andreas <aw...@sw...>]

Hi Kulwant

On Sat, 2003-12-06 at 22:46, Kulwant Bhogal wrote:

> Hello all,
> 
> My question: 
> 
> Which is the best version of the 
> 
> kernel        
> boothack

Use the latest boothack available.

> but kernel panics when initialising
> the PCMCIA ethernet card.

Hmm, hasn't there been a problem in recent kernels with apne? I think it
got fixed in kernel 2.4.21 (search the apus devel archive for thread
"Kernel 2.4.21". The posts are from july/august 2003.).

If it doesn't work at all, remove the pcmcia network card. Though I
guess you want some kind of ethernet connection.. ;)

> With some other version it panic'd when trying to set the
> h/w clock.

Don't know about this one.

-- 
Best wishes,
Andi

Re: Version(s) working with BlizzardPPC

[Kulwant B. <kul...@bt...>]

Hello Alan, 

Thanks for your reply. I will try the versions you suggested. I have already
tried kernel 2.4.20 with the latest (010623) boothack but that panics when
starting (after installation) when it tries to set the hardware clock. And
yes I did download the above from the sourceforge site. 

I did find your reply a little confusing as within the same paragraph you
say "apart from the SCSI ...... .... works fine", whilst in the next sentence you say
"scsi driver is built in by default". 

From that I imagine three posibilities: 

1. You meant the SCSI support is built in and not by some external
module/driver. 

2. SCSI is not supported by 2.2.13. or 

3. I haven't got a clue what you meant.  

In any case I will try that version. 

Is someone working on 2.4.23? I do find it strange that an older version is
more compatible than a new version. I also wonder how many versions of the
BlizzardPPC there are to make these compatibility problems manifest
themselves - I mean, surely a hardware revision wouldn't implement a
different way of setting the hardware clock - or does that question simply
reveal that I haven't got a clue about what I am talking about and know very
little about a) what setting the hardware clock actually means and b)
haven't a clue about what is involved in such a procedure. In my simple
(unexperienced and sometimes clueless) mind you'd poke a few bytes into the
appropriate locations (which has just made me realise that the locations could
have changed with different revisions of board.....) and hey presto.....

Is there a way of determining what version of board I have (without taking
the Tower apart again)? 

A "Report a problem" selection from the Installer program gave some output
which said my PPC processor was a "603ev". Does that mean "evaluation 
version" or something? Or am I clutching at straws? 

Kind regards, 

Kulwant


-----Original Message-----
From:  [mailto:a.l...@lb...]
Sent: 08/12/2003 13:18:20
To: 
Subject: Re: Version(s) working with BlizzardPPC


> Hi,

>> Which is the best version of the 
>> 
>> kernel        
>> boothack

> for kernel, its very dependant on hardware. I've found the best general
> A1200 one is the 2.4.13 dated  20011030  you might have joy with 2.4.17,
> or even the very latest. there are some issues that need to be ironed out
> for a nice 2.4.23 
> for boothack, the VERY LATEST. (010623)

> these files are both available from the Linux-APUS sourceforge site
> (official homepage) 
> eg

> http://linux-apus.sourceforge.net/kernels/

> and

> http://sourceforge.net/project/showfiles.php?group_id=5907

> DONT 'save as' images from this page, as they are links to a redirector
> site! 
>> A1200 with BlizzardPPC with 35Gb SCSI storage (& 420Mb IDE), 128Mb RAM, 
>> BVision Permidia2 3DGFX (8Mb RAM)
>> PCMCIA Ethernet card. 

> yep. apart from the SCSI, the 2.4.13 i mentioned above works on that
> system fine. pm2fb driver for the BVision, APNE driver module for the
> PCMCIA network card, scsi driver is built in by default 
> alan
> 

Version(s) working with BlizzardPPC

[Kulwant B. <kb...@bl...>]

Hello all,

My question: 

Which is the best version of the 

kernel        
boothack

and anything else 

I need to get LinuxPPC running on my machine: 

A1200 with BlizzardPPC with 35Gb SCSI storage (& 420Mb IDE), 128Mb RAM, 
BVision Permidia2 3DGFX (8Mb RAM)
PCMCIA Ethernet card. 

Please be specific with the versions of each and if possible where to get
each from. The stuff from http://ftp.no.debian.org//debian/dists/woody/main/disks-powerpc/current/apus/
has an old boothack which complains about "compressed image too large" and
the one from sourceforge starts booting but kernel panics when initialising
the PCMCIA ethernet card. With some other version it panic'd when trying to set the
h/w clock. 

Kind regards, 

Kulwant

pb with a4091

[Wilfried C. <wca...@no...>]

Hi,

i want to use a cdrom connected to my a4091 scsi card. To do this, i =
compiled the kernel with the a4091 support actived. But at boot time, i =
have a kernel panic when detecting the a4091 card. Is this a pb with the =
driver ? Do i have to active other thing when compiling the kernel ?

Thanks, Willy.

Re: Debian refuses to boot!

[Daniel I. R. <ir...@ip...>]

Hello a.l...@lb...

> please add  console=tty0   to the end of your bootline...i'd also
> suggest a visit to: http://linux-apus.sourceforge.net/FAQ.html

     I don't know what happened, my installation was running and booting
after putting console=tty0 as you suggested, but .. now I get the same
problem again: (Something with the keyb) :-( (see attached dmesg output
file)

I'm using this to boot:

(see attached boot file)

any ideas?

Thanks in advance,

Daniel

Re: Debian refuses to boot!

[Daniel I. R. <ir...@ip...>]

Hello again,

I've successfully installed Debian 3.0 woody PPC on my Amiga 1200  thanks to
all of you who helped me  :-)

I updated the sources list with this command (according to the readme in the
2nd binary CD of Debian distro):

apt-cdrom add

     then did this:

apt-get update

     and then I was able to install XWindow using *dselect*

This added the list of installable packages of the 3 first binary CDs using
*dselect*

     I want to get Xwindow working with KDE, so I'll install KDE too but
could anyone help me with XWindow system? Do I need any specific amiga
files to get it working? How can I configure it? If anyone can provide me
with a sample (but working) Xfree86config prefs file for BVision I will
really appreciate it.

best regards,

Daniel