Boot, cpu, testing

OK I'm going to install the last build on my iBook G4 (the
one without
BlueTooth). I already try a few times some weeks ago and I
have a few
questions before beginning:

- I read on the mailing lists here[1][2] some interesting
things about
   booting NetBSD directly ; but since I wasn't here at the
time I may
   not fully understand what they're trying to accomplish.
So here's my
   question: can NetBSD boot from an iBook G4 with OFW3
without having
   to type boot in OFW?

[1] http://mail-index.netbsd.org/port-macppc/2001/09/02
/0002.html
[2] http://mail-index.netbsd.org/port-macppc/2002/09/19
/0001.html

- Last time I installed it NetBSD reported a 600 Mhz CPU (it
is a
   1.2 Ghz actually). I think OFW don't report the correct
values,
   but is there a way to change the frequency after boot?

- Do you want to test something on this machine? (I already
know I
   have to test radeonfb). The details of the machine are
below.

Plus, I may help trying to fix or improve things. Is there a
place
where I can see what needs to be done?



Machine details:
----------------

Hardware:

     Hardware Overview:

       Machine Model: iBook G4
       CPU Type: PowerPC G4  (1.2)
       Number Of CPUs: 1
       CPU Speed: 1.33 GHz
       L2 Cache (per CPU): 512 KB
       Memory: 512 MB
       Bus Speed: 133 MHz
       Boot ROM Version: 4.8.7f1
       Serial Number: 4H5050SLS88

Network:

     Internal Modem:

       Interface: modem
       Type: PPP (PPPSerial)

     Built-in Ethernet:

       Interface: en0
       Type: Ethernet
       Ethernet Address: 00:11:24:73:2e:d6

     AirPort:

       Interface: en1
       Type: AirPort
       Ethernet Address: 00:11:24:2d:39:ac

Memory:

     DIMM0/BUILT-IN:

       Size: 256 MB
       Type: Built-in
       Speed: Built-in

     DIMM1/J7:

       Size: 256 MB
       Type: DDR SDRAM
       Speed: PC2700U-25330

PCI/AGP Cards:

     ATY,RV280M9+:

       Type: display
       Bus: AGP
       VRAM (Total): 32 MB
       Vendor: ATI (0x1002)
       Device ID: 0x5c63
       Revision ID: 0x0001
       ROM Revision: 113-xxxxx-142

         Display:

           Type: display
           Display Type: LCD
           VRAM (In Use): 32 MB
           Resolution: 1024 x 768
           Depth: 32-bit Color
           Main Display: Yes
           Mirror: Off
           Online: Yes

         ATY,Via_B:

           Status: No display connected

ATA:

     ATA-3 Bus:

         MATSHITA CD-RW CW-8123:

           Manufacturer: MATSHITA
           Model: MATSHITA CD-RW CW-8123
           Revision: CA10
           Serial Number:
           Drive Type: CD-RW/DVD-ROM
           Disc Burning: Apple Supported/Shipped
           Removable Media: Yes
           Detachable Drive: No
           Protocol: ATAPI
           Unit Number: 0
           Socket Type: Internal

     ATA-6 Bus:

       Vendor ID: 0x106b
       Device ID: 0x003b
       Revision ID: 0x0000

         FUJITSU MHT2060AT:

           Capacity: 55.89 GB
           Model: FUJITSU MHT2060AT
           Revision: 8164
           Serial Number: NN94T4C145SU
           Removable Media: No
           Detachable Drive: No
           BSD Name: disk0
           Protocol: ATA
           Unit Number: 0
           Socket Type: Internal
           OS9 Drivers: No

             Untitled:

               Capacity: 55.76 GB
               Available: 47.57 GB
               Writable: Yes
               File System: Journaled HFS+
               BSD Name: disk0s3
               Mount Point: /

USB:

     USB Bus:

       Vendor Name: Apple Computer, Inc.
       Speed: Up to 12 Mb/sec
       Product ID: 32773 ($8005)
       Bus Power (mA): 500

     USB Bus:

       Vendor Name: Apple Computer, Inc.
       Speed: Up to 12 Mb/sec
       Product ID: 32773 ($8005)
       Bus Power (mA): 500

     USB Bus:

       Vendor Name: Apple Computer, Inc.
       Speed: Up to 12 Mb/sec
       Product ID: 32773 ($8005)
       Bus Power (mA): 500

     USB High-Speed Bus:

       Vendor Name: Apple Computer, Inc.
       Speed: Up to 480 Mb/sec
       Product ID: 32774 ($8006)
       Bus Power (mA): 500

FireWire:

     FireWire Bus:

       Speed: 400 Mb/sec Speed

AirPort Card:

     AirPort Card Information:

       Wireless Card Type: AirPort Extreme
       Wireless Card Locale: Worldwide
       Wireless Card Firmware Version: 3.4.3f1 (3.50.37.p4)

Modems:

     Modem Information:

       Driver: MotorolaSM56K.kext (v1.2.0)
       Hardware Version: Version 1.0
       Interface Type: I2S
       Country: B5 (United States, Latin America)
       Modem Model: Jump
       Modulation: V.92

-- 
Loïc Hoguin
Dev:Extend

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On Mar 15, 2007, at 16:37, Loic Hoguin wrote:

> OK I'm going to install the last build on my iBook G4
(the one without
> BlueTooth). I already try a few times some weeks ago
and I have a few
> questions before beginning:
>
> - I read on the mailing lists here[1][2] some
interesting things about
>   booting NetBSD directly ; but since I wasn't here at
the time I may
>   not fully understand what they're trying to
accomplish. So here's my
>   question: can NetBSD boot from an iBook G4 with OFW3
without having
>   to type boot in OFW?

Maybe, maybe not. I never tried it but newer OF has a
tendency to omit 
initializing the console when booting without user
intervention. There 
are tricks to get around that though.

> - Last time I installed it NetBSD reported a 600 Mhz
CPU (it is a
>   1.2 Ghz actually). I think OFW don't report the
correct values,
>   but is there a way to change the frequency after
boot?

It's the other way around - NetBSD actually measures CPU
speed and 
reports what it finds without looking at OF ( which will
very likely 
report 1.2GHz ). There are several ways Apple used to switch
CPU speed, 
I have a patch to support one of them ( so I can switch my
800MHz iBook 
G4 between 600 and 800MHz ) - this may or may not work on
yours. I'll 
clean it up and commit.

> - Do you want to test something on this machine? (I
already know I
>   have to test radeonfb). The details of the machine
are below.

Please mail me the output of dmesg and ofctl -p - I need to
verify a 
few things, namely ADB attachment and CPU speed control.

> Plus, I may help trying to fix or improve things. Is
there a place 
> where I can see what needs to be done?

hmm, in no particular order ( all -current ):
- - please check if the pmu, nadb, adbkbd etc. drivers do
the right thing
- - does backlight control via wsconsctl work correctly?
- - what kind of thermal monitor / fan controller does your
iBook use? 
Usually there's a 'fan' node hanging on one of the i2c
buses, the 
'compatible' property should give a hint what kind of chip
it is. If 
it's unsupported you may want to dig for a datasheet and
write a driver 
( easy enough - see src/sys/dev/i2c/adt7647.c and adm1030.c
for 
examples )
What we're really lacking is decent power management support
though.
Another thing would be to find out how your iBook switches
CPU speed. 
Mine uses two gpios on mac-io for this - one switches the
bus clock, 
another one the core voltage. Both are easily identified by
their 
names.

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iQEVAwUBRfnHB8pnzkX8Yg2nAQI8DwgAuEJ9sIGadEth05nVsDrrl3wr7S/L
1slA
ZVr/m9HLOxeEb+bdSW6J4kKOt1dPDLbYzyP4pySuGu7v5hy0kM7+PAMnF6Nj
vpk3
BQUe65KswkQlvtj/Fc/QbQM9DOiJ1v23IH9gvWUEOdC8tES0NMpjfVLtM3mz
hqQ8
IJVTfbp6uPQxlinpgvZaD3IWyg+SiWFwVkk1eBeKBeJeE7QtUhS5aYSi2+bx
1DnV
hg4BMgH/gU+LLsKea7/pLtTojngCjaWQIRVUWocr2aHDwH85xwwJ05o7Vklx
EAJc
wWw0cVL3Rp2afp+V3g7dYbxJzZbenU5E9Cs7XwfQsbV/Cthb5FJs0Q==
=20dV
-----END PGP SIGNATURE-----

On Fri, 16 Mar 2007 00:24:05 +0100, Magnus Henoch
<mangefreemail.hu>  
wrote:

> "Loic Hoguin" <essendev-extend.eu> writes:
>
>> - I read on the mailing lists here[1][2] some
interesting things about
>>   booting NetBSD directly ; but since I wasn't here
at the time I may
>>   not fully understand what they're trying to
accomplish. So here's my
>>   question: can NetBSD boot from an iBook G4 with
OFW3 without having
>>   to type boot in OFW?
>
> Yes.  I set boot-device to 'hd:3,ofwboot.xcf',
boot-file to
> 'hd:5,/netbsd', and boot-command to '"
screen" output boot'.  The
> first path refers to ofwboot.xcf in the root of my Mac
OS partition,
> and the second to the kernel in my NetBSD root
partition.  The third
> setting is needed to initialize the display.

Thanks for the tip, it worked. I guess I was missing the
'" screen" output'
part last time. I made a 32MB HFS partition with only
ofwboot.xcf at hd:2,
a swap and netbsd partition. And it works like a charm.

-- 
Loïc Hoguin
Dev:Extend

On Thu, 15 Mar 2007 23:21:59 +0100, Michael Lorenz
<macallannetbsd.org>  
wrote:

>> - Do you want to test something on this machine? (I
already know I
>>   have to test radeonfb). The details of the
machine are below.
>
> Please mail me the output of dmesg and ofctl -p - I
need to verify a few  
> things, namely ADB attachment and CPU speed control.

I just got some weird bugs trying to install current.

First the ./build.sh -O ../obj -T ../tools install=/ failed.
It said
"Bus error (core dumped)", but the system was then
unusable and I couldn't
reboot (freezed at the part "root file system type:
ffs", the last line
outputed by the kernel at boot). I had to reinstall sets
from my netbsd-4
CD to be able to boot. I'm going to retry the install now.

Next when rebooting later I had a screen with (it seems) a
wrong h/vsync,
just after radeonfb is initialized. I couldn't really see
much with all
the flickering. Next reboot was OK though.

I'm going to continue testing.

-- 
Loïc Hoguin
Dev:Extend

On Mon, 19 Mar 2007 21:14:06 +0100, Loic Hoguin
<essendev-extend.eu>  
wrote:

> Next when rebooting later I had a screen with (it
seems) a wrong h/vsync,
> just after radeonfb is initialized. I couldn't really
see much with all
> the flickering. Next reboot was OK though.
>
> I'm going to continue testing.

The problem at boot occured with a current kernel, I didn't
have this with
the netbsd-4 (14 march) one (but that doesn't mean the
problem wasn't there
I guess).

A few more details:

- I can switch to other screens, however there's only a
blank screen there.
- I can start X, but can't switch to other screens while in
X.
- When X closes, the display mode doesn't seem to be
changed, and the  
result
   is a garbled screen
- wsconsctl works for backlight, although I have the values
reversed (0 is
   brighter, 255 is off) - btw the display is off starting
with a value of  
220
   (approximately, and not always the same value...)
   - the values were in the reverse order with the netbsd4
kernel (0 off)

Regards,

-- 
Loïc Hoguin
Dev:Extend

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On Mar 20, 2007, at 16:51, Loic Hoguin wrote:

> On Thu, 15 Mar 2007 23:21:59 +0100, Michael Lorenz 
> <macallannetbsd.org> wrote:
>
>> - - please check if the pmu, nadb, adbkbd etc.
drivers do the right 
>> thing
>
> OK so I compiled a kernel with support for these
devices, and I got a 
> panic
> caused by the adbkbd driver. Sources are current from
some minutes ago.
>
> I pressed Command+F1 and it outputted this:
>
> adbkbd0: power button pressed, shutting down!
> syncing disks... trap: kernel read DSI trap 
0x7fa00026 by 0x3b247c 
> (DSISR 0x40000000, err=14), lr 0x38bfcc
> panic: trap
> db >

Please try that with powerd running.
Anyway, this should happen only when you press the power
button.

No need for a trace, I know where that comes from. Sure you
didn't 
touch the power button?

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iQEVAwUBRgDLxMpnzkX8Yg2nAQL6EQf/WItNYRbBE/uOqq0IT9M/vi4lHaSo
SzoY
jgo0gC6O7lJkTe/ZulvGSliOBPW0u3K3ZGi9Ju25Znvnw7tT57CNBveoRwMG
5GtS
BbOT/0HGK82kSi5uThrkzRD6BSwHOmLG85EuvX/EgVuzh9c1nzLgDGKQyx36
m3V1
VYbNXrGBwcNTgyJLzQJWR3vTxCQTCWlJBfO++WN9y5XO2KpYu18/bK+Ydz0P
1P93
E2YbApO20zhfvs92hBRzJZFCFF5h6r+32AbAlhetzp6fX68YanRwbtl8zpPd
TIKY
A2P2iUa+tIQWHAaSYHBqFAgiEvSLRI/G6Sk6GbJ0v1nFwKipNJOevw==
=yJMN
-----END PGP SIGNATURE-----

On Wed, 21 Mar 2007 07:08:03 +0100, Michael Lorenz
<macallannetbsd.org>  
wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hello,
>
> On Mar 20, 2007, at 16:51, Loic Hoguin wrote:
>
>> On Thu, 15 Mar 2007 23:21:59 +0100, Michael Lorenz 

>> <macallannetbsd.org> wrote:
>>
>>> - - please check if the pmu, nadb, adbkbd etc.
drivers do the right  
>>> thing
>>
>> OK so I compiled a kernel with support for these
devices, and I got a  
>> panic
>> caused by the adbkbd driver. Sources are current
from some minutes ago.
>>
>> I pressed Command+F1 and it outputted this:
>>
>> adbkbd0: power button pressed, shutting down!
>> syncing disks... trap: kernel read DSI trap 
0x7fa00026 by 0x3b247c  
>> (DSISR 0x40000000, err=14), lr 0x38bfcc
>> panic: trap
>> db >
>
> Please try that with powerd running.

I will when I'm back from work.

> Anyway, this should happen only when you press the
power button.
>
> No need for a trace, I know where that comes from. Sure
you didn't touch  
> the power button?

No, I didn't. Didn't try the power button yet.

Regards,

-- 
Loïc Hoguin
Dev:Extend

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On Mar 21, 2007, at 04:22, Loic Hoguin wrote:

>> Anyway, this should happen only when you press the
power button.
>>
>> No need for a trace, I know where that comes from.
Sure you didn't 
>> touch the power button?
>
> No, I didn't. Didn't try the power button yet.

That's odd. Please send me the dmesg output for this kernel,
your 
keyboard probably needs just another special case ( so far I
know of 
three different scancodes sent by power buttons on various
ADB 
keyboards - why couldn't they use the same everywhere? )
The old driver doesn't have this problem because it doesn't
know or 
care about the power button.

have fun
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iQEVAwUBRgFUW8pnzkX8Yg2nAQK8pAgAiVh9Fe6iDfY2UWwK00hTucwI70ge
d0pq
w/Yi+vT+SeBG7P35q1dETNXGrwBqUANJfITILh09jp8pYttLCVumehPyL4CX
+Fe9
dS3/KF6hXSh8DvcRU33OoPaiLW3ml3RvLU3t6+HT+PgnoCT8fVH0Ug69eoYL
piHc
vAONICYgl/ggU4zVlt3VlsTmO0zVrPomOy9PLZyewg0vnBI1DNjO7LTtCOP5
YT+D
CM3DjZVpUZr+f0PdR7zlDATyrmFISRoG+tqYuF4EN9qQ0h4t865gFNObezh/
jnJa
XMZsQj90k3lViCUzm5BOw3kcmQs5itg7UDA2Kj2Bs2uP2aOVEBKg2Q==
=2ags
-----END PGP SIGNATURE-----

On Wed, 21 Mar 2007 16:50:51 +0100, Michael Lorenz
<macallannetbsd.org>  
wrote:

> Hello,
>
> On Mar 21, 2007, at 04:22, Loic Hoguin wrote:
>
>>> Anyway, this should happen only when you press
the power button.
>>>
>>> No need for a trace, I know where that comes
from. Sure you didn't  
>>> touch the power button?
>>
>> No, I didn't. Didn't try the power button yet.
>
> That's odd. Please send me the dmesg output for this
kernel, your  
> keyboard probably needs just another special case ( so
far I know of  
> three different scancodes sent by power buttons on
various ADB keyboards  
> - why couldn't they use the same everywhere? )
> The old driver doesn't have this problem because it
doesn't know or care  
> about the power button.

Hello,

The dmesg output is attached.

-- 
Loïc Hoguin
Dev:Extend

Hello,

>> Please try that with powerd running.
>
> I will when I'm back from work.

With powerd the power button shutdown successfully, and the
Fn+Command keys shutdowns too. There's no panic.

>> Anyway, this should happen only when you press the
power button.
>>
>> No need for a trace, I know where that comes from.
Sure you didn't  
>> touch the power button?
>
> No, I didn't. Didn't try the power button yet.

With powerd stopped the power button and the Fn+Command do
panic
in ffs_sync (more precisely:
specificdata_getspecific_unlock).

All with 4.99.16.

-- 
Loïc Hoguin
Dev:Extend