linux-apus-devel

2.3 and ami_int3

[Nicholai B. <nic...@ch...>]

Hello,

I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and the
builtin serial driver didn't compile at all. The pm2 driver still needs ami_int3 so
I put it back. Should it be there or not?

I still didn't manage to boot all the way since I had a lockup while
mounting filesystems. I will look into all of this tomorrow. Just wanted to
let you know.

Friendly,

Nicholai 

Re: 2.3 and ami_int3

[<fp...@zu...>]

This is a no-work report as well. I managed to compile and boot, but after
the screen goes black, I see exactly one blink of the cursor in the top
left and exactly one heartbeat blink before it freezes.

Machine: A4k, CV64/3D, IDE HD. I compiled everything I could as modules.
-- 
Frank Petzold, IBM Zurich Research Laboratory, Säumerstrasse 4,
CH-8803 Rüschlikon/Switzerland, Tel.  +41-1-724-84-42  Fax.   +41-1-724-89-56
 Business email: fp...@zu...          Private email: pe...@he...
   The opinions expressed here are mine and not necessarily those of IBM.

Re: 2.3 and ami_int3

[Michel <da...@re...>]

Frank Petzold wrote:
> 
> This is a no-work report as well. I managed to compile and boot, but after
> the screen goes black, I see exactly one blink of the cursor in the top
> left and exactly one heartbeat blink before it freezes.

Nothing from dmesg?


Michel


PS: Nicholai: Wouldn't it be a good idea to include dmesg in the boothack
archive?

-- 
...and that is how we know the Earth to be banana-shaped.
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Frank P. <fp...@zu...>]

On Wed, May 31, 2000 at 08:36:22AM +0200, Michel Dänzer wrote:
> Frank Petzold wrote:
> > 
> > This is a no-work report as well. I managed to compile and boot, but after
> > the screen goes black, I see exactly one blink of the cursor in the top
> > left and exactly one heartbeat blink before it freezes.
> 
> Nothing from dmesg?

I actually had debugging switched off. And no time for a second try.
-- 
Frank Petzold, IBM Zurich Research Laboratory, Säumerstrasse 4,
CH-8803 Rüschlikon/Switzerland, Tel.  +41-1-724-84-42  Fax.   +41-1-724-89-56
 Business email: fp...@zu...          Private email: pe...@he...
   The opinions expressed here are mine and not necessarily those of IBM.

Re: 2.3 and ami_int3

[Frank P. <fp...@zu...>]

On Wed, May 31, 2000 at 08:36:22AM +0200, Michel Dänzer wrote:
> Frank Petzold wrote:
> > 
> > This is a no-work report as well. I managed to compile and boot, but after
> > the screen goes black, I see exactly one blink of the cursor in the top
> > left and exactly one heartbeat blink before it freezes.
> 
> Nothing from dmesg?

From what I remember (did the test yesterday) ....234567K<<<<<<<<<<<<
-- 
Frank Petzold, IBM Zurich Research Laboratory, Säumerstrasse 4,
CH-8803 Rüschlikon/Switzerland, Tel.  +41-1-724-84-42  Fax.   +41-1-724-89-56
 Business email: fp...@zu...          Private email: pe...@he...
   The opinions expressed here are mine and not necessarily those of IBM.

Re: 2.3 and ami_int3

[Michel <dae...@st...>]

Frank Petzold wrote:

> > > This is a no-work report as well. I managed to compile and boot, but
> > > after the screen goes black, I see exactly one blink of the cursor in
> > > the top left and exactly one heartbeat blink before it freezes.
> >
> > Nothing from dmesg?
> 
> From what I remember (did the test yesterday) ....234567K<<<<<<<<<<<<

That looks rather like bootmesg to me...


Michel


-- 
0 and 1. Now what could be so hard about that?
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Frank P. <fp...@zu...>]

On Mon, Jun 05, 2000 at 10:06:56AM +0200, Michel Dänzer wrote:
> Frank Petzold wrote:
> 
> > > > This is a no-work report as well. I managed to compile and boot, but
> > > > after the screen goes black, I see exactly one blink of the cursor in
> > > > the top left and exactly one heartbeat blink before it freezes.
> > >
> > > Nothing from dmesg?
> > 
> > From what I remember (did the test yesterday) ....234567K<<<<<<<<<<<<
> 
> That looks rather like bootmesg to me...

Correct.
-- 
Frank Petzold, IBM Zurich Research Laboratory, Säumerstrasse 4,
CH-8803 Rüschlikon/Switzerland, Tel.  +41-1-724-84-42  Fax.   +41-1-724-89-56
 Business email: fp...@zu...          Private email: pe...@he...
   The opinions expressed here are mine and not necessarily those of IBM.

Re: 2.3 and ami_int3

[Michel <da...@re...>]

Nicholai Benalal wrote:

> I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and the
> builtin serial driver didn't compile at all.

Have you noticed that it's not the same serial driver as in 2.2?

I investigated a bit and found the fields it complains about (xmit_*) in
struct gs_port (include/linux/generic_serial.h).


Michel


-- 
True programmers never die, they just branch to an odd address.
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Roman Z. <zi...@fh...>]

Hi,

> I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and the
> builtin serial driver didn't compile at all. The pm2 driver still needs ami_int3 so
> I put it back. Should it be there or not?

I tested only amifb and it compiled. Which gcc are you using?

bye, Roman

Re: 2.3 and ami_int3

[Nicholai B. <nic...@ch...>]

Hello Roman

Den 31-Maj-00, skrev Roman Zippel:

RZ> Hi,
RZ> 
RZ>> I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and
RZ>> the builtin serial driver didn't compile at all. The pm2 driver still
RZ>> needs ami_int3 so I put it back. Should it be there or not?
RZ> 
RZ> I tested only amifb and it compiled. Which gcc are you using?
RZ> 
RZ> bye, Roman
RZ> 

I'm using gcc-2.95.2-1g but the error about too many arguments (below) can't
have anything to do with my compiler, can it? I'm using the 2.3 includes.

amifb.c: In function `amifb_interrupt':
amifb.c:1930: variable `server' has initializer but incomplete type
amifb.c:1930: warning: excess elements in struct initializer
amifb.c:1930: warning: (near initialization for `server')
amifb.c:1930: warning: excess elements in struct initializer
amifb.c:1930: warning: (near initialization for `server')
amifb.c:1930: storage size of `server' isn't known
amifb.c:1974: too many arguments to function `amiga_do_irq_list' 

Btw, any comment on ami_int3 (see amiints.c)? To be or not to be? Jesper
wasn't sure about this.

/Nicholai

Re: 2.3 and ami_int3

[Roman Z. <zi...@fh...>]

Hi,

> I'm using gcc-2.95.2-1g but the error about too many arguments (below) can't
> have anything to do with my compiler, can it? I'm using the 2.3 includes.

Oh, that's my fault, I didn't recomile the whole kernel...
I committed a change, but you can also simply remove that last argument (I
don't know when the switch at the cvs server to Jesper's tree will
happen).

> Btw, any comment on ami_int3 (see amiints.c)? To be or not to be? Jesper
> wasn't sure about this.

I'll clean up amifb, so it will go through ami_int3().

bye, Roman

Re: 2.3 and ami_int3

[Michel <dae...@st...>]

Nicholai Benalal wrote:

> I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and the
> builtin serial driver didn't compile at all. The pm2 driver still needs
> ami_int3 so I put it back. Should it be there or not?

It still doesn't boot with pm2fb here, I get tons of 'amiga_disable_irq:
trying to disable Amiga auto IRQ 3' messages.

Is that related? What was the fix/workaround?


Michel


-- 
I believe the technical term is "Oops!"
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Nicholai B. <nic...@ch...>]

Hello Michel

Den 03-Jun-00, skrev Michel Dänzer:

MD> Nicholai Benalal wrote:
MD> 
MD>> I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and
MD>> the builtin serial driver didn't compile at all. The pm2 driver still
MD>> needs ami_int3 so I put it back. Should it be there or not?
MD> 
MD> It still doesn't boot with pm2fb here, I get tons of 'amiga_disable_irq:
MD> trying to disable Amiga auto IRQ 3' messages.
MD> 
MD> Is that related? What was the fix/workaround?
MD> 

With Jesper's latest 2.3 kernel I remember doing free_irq(IRQ_AMIGA_AUTO_3, NULL); 
before the request_irq(). That should work allthough I don't think it's a
proper fix. Another way to avoid this is to not include amifb. I'm waiting
for someone to find out how this should really be handled.

Btw, 2.3 seems to works for me now except for mounting affs partitions.
mount never returns when I try that.

/Nicholai

Re: 2.3 and ami_int3

[Michel <dae...@st...>]

Nicholai Benalal wrote:
> 
> Hello Michel
> 
> Den 03-Jun-00, skrev Michel Dänzer:
> 
> MD> Nicholai Benalal wrote:
> MD>
> MD>> I compiled linux-2.3 from the CVS. A lot of drivers such as amifb and
> MD>> the builtin serial driver didn't compile at all. The pm2 driver still
> MD>> needs ami_int3 so I put it back. Should it be there or not?
> MD>
> MD> It still doesn't boot with pm2fb here, I get tons of 'amiga_disable_irq:
> MD> trying to disable Amiga auto IRQ 3' messages.
> MD>
> MD> Is that related? What was the fix/workaround?
> MD>
> 
> With Jesper's latest 2.3 kernel I remember doing free_irq(IRQ_AMIGA_AUTO_3,
> NULL); before the request_irq().

Where - in amifb.c, pm2fb.c or fbmem.c?


> That should work allthough I don't think it's a proper fix. Another way to
> avoid this is to not include amifb.

I only built amifb as a module, will test without it altogether...


> I'm waiting for someone to find out how this should really be handled.

I join you :)


> Btw, 2.3 seems to works for me now except for mounting affs partitions.
> mount never returns when I try that.

Has AFFS already been fixed in 2.3.99-pre9?


Michel


-- 
The computer revolution is over. The computers won.
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Nicholai B. <nic...@ch...>]

Hello Michel

Den 03-Jun-00, skrev Michel Dänzer:


MD>> With Jesper's latest 2.3 kernel I remember doing
MD>> free_irq(IRQ_AMIGA_AUTO_3, NULL); before the request_irq().
MD> 
MD> Where - in amifb.c, pm2fb.c or fbmem.c?

As far as I can remember it was in amifb.c. Just check where the
request_irq() fails and insert the free_irq() call before that.

MD>> Btw, 2.3 seems to works for me now except for mounting affs partitions.
MD>> mount never returns when I try that.
MD> 
MD> Has AFFS already been fixed in 2.3.99-pre9?
MD> 

I thought so but I'm not sure.

/Nicholai

Re: 2.3 and ami_int3

[Roman Z. <zi...@fh...>]

Hi,

> With Jesper's latest 2.3 kernel I remember doing free_irq(IRQ_AMIGA_AUTO_3, NULL); 
> before the request_irq(). That should work allthough I don't think it's a
> proper fix. Another way to avoid this is to not include amifb. I'm waiting
> for someone to find out how this should really be handled.

If you don't compile in amifb, you probably need to enable ami_int3 in
arch/ppc/amiga/amiints.c in the amiga_default_handler array. I have a new
amifb that should fix this, but I'd like to test that a bit more (it's
ages ago I worked on that driver :) ) and I'm currently away from home, so
new patches from me only somewhere next week.

bye, Roman

Re: 2.3 and ami_int3

[Roman Z. <zi...@fh...>]

Hi,

> It still doesn't boot with pm2fb here, I get tons of 'amiga_disable_irq:
> trying to disable Amiga auto IRQ 3' messages.

This shouldn't happen anymore with my latest cvs changes (at least it
went away for me... :) ), IMO it was a reentrancy problem for shared
interrupts.

bye, Roman

Re: 2.3 and ami_int3

[Michel <da...@re...>]

Roman Zippel wrote:
> 
> Hi,
> 
> > It still doesn't boot with pm2fb here, I get tons of 'amiga_disable_irq:
> > trying to disable Amiga auto IRQ 3' messages.
> 
> This shouldn't happen anymore with my latest cvs changes (at least it
> went away for me... :) ), IMO it was a reentrancy problem for shared
> interrupts.

Hmm... my tree was up to date - maybe I have confused the kernel images...
I'll have another try next weekend.


Michel


PS: When I just did another update, I got lots of messages like this:

cvs server: mtio.h: 1.1.1.1 and working file are different file types
cvs server: mtio.h: hard linkage of 1.1.1.1 and working file do not match
M include/linux/mtio.h

Is this related to the "PreservePermissions" option in CVSROOT/config (which
doesn't help for the commit notices) ?

-- 
Why drink & drive when you can smoke and fly???
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Michel <dae...@st...>]

Roman Zippel wrote:
> 
> Hi,
> 
> > It still doesn't boot with pm2fb here, I get tons of 'amiga_disable_irq:
> > trying to disable Amiga auto IRQ 3' messages.
> 
> This shouldn't happen anymore with my latest cvs changes (at least it
> went away for me... :) ), IMO it was a reentrancy problem for shared
> interrupts.

I still have the problem, whether I build amifb in the kernel, as a module or
not at all.


Michel


-- 
It's not a bug, it's tradition!
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS

Re: 2.3 and ami_int3

[Roman Z. <zi...@fh...>]

Hi,

> I still have the problem, whether I build amifb in the kernel, as a module or
> not at all.

What is normally in your /proc/interrupts?
I had the same problem, after inserting the ariadne2 module and the
first few interrupts the system was dead. After I cleaned up a bit the
interrupt stuff it went away. Could you insert some prints in
enable_irq/disable_irq (and maybe also a check for in_interrupt())?

bye, Roman

Re: 2.3 and ami_int3

[Michel <dae...@st...>]

Roman Zippel wrote:

> > I still have the problem, whether I build amifb in the kernel, as a module
> > or not at all.
> 
> What is normally in your /proc/interrupts?
> I had the same problem, after inserting the ariadne2 module and the
> first few interrupts the system was dead. After I cleaned up a bit the
> interrupt stuff it went away. Could you insert some prints in
> enable_irq/disable_irq (and maybe also a check for in_interrupt())?

I'm away from my Amiga until tomorrow at least, so I hope someone else can
jump in.


Michel


-- 
Therapy is expensive, popping bubble wrap is cheap. You choose.
______________________________________________________________________________
Earthling Michel Dänzer (MrCooper)  \  CS student and free software enthusiast
Debian GNU/Linux (powerpc,i386) user \   member of XFree86, Team *AMIGA*, AUGS