difference cdrom boot and 9fat boot?

difference cdrom boot and 9fat boot?

[staalmannen]

Dear,

I just made a native install of 9Front (nightly build from yesterday).
The live CD works
flawlessly and the install procedure runs without a hickup. However,
when booting from the livecd I get a system memory of 2047MB whereas
booting from my 9fat partition (via GRUB2) gives me a memory of 8MB
(1MB kernel, 6MB user, 39MB swap)
just before it crashes.

I had the same issue with a previous 9atom install, but there I could
resolve the issue by copying the 9pcf image from the official plan9
iso (which is a bit bigger: 3,4MB compared to 9atom/9front ~2,6MB) to
9fat. This solution will however not work with 9Front since it has a
different root file system. I found 9Front because I was searching for
a broadcom ethernet driver for Plan9 and ended up at 9Front (this
driver works great (on the live CD) btw!).  I checked a diff between
the official pcf and the 9front pcf configuration files. As far as I
can see, there are no obvious differences there that could explain why
the official 9pcf can detect the memory and the 9atom/9front 9pcf:es
can not (and also not why the official 9pcf is so much bigger).

What do you guys suggest? Should I revert to my 9atom/plan9 setup and
just copy the broadcom driver + the official plan9.iso pcf
configuration file for a recompile of the kernel or should I try to
solve the issue with the 9Front boot?

I have been editing my plan9.ini to avoid weird BIOS-related issues
with memory detection.
The weird thing though is that it works on the live CD. Do you guys
know what kind of black magick the cd boot might do that the 9fat boot
does not do?

For those interested, here is documentation of the hardware I am using
(yeah I know. Fedora might suck but I need it at work):
http://www.smolts.org/client/show/pub_e908de8a-222b-4ffe-88c3-d5e66d7ae61e

Ps. Another question: The install procedure makes a small swap
partition in the Plan9 partition. Is it possible to use a Linux swap
partition for Plan9 instead? Ds.

Re: difference cdrom boot and 9fat boot?

[cinap_lenrek]

this is really wired... the iso and fat bootloaders (we
dont use 9load) share most of the code except filesystem
specific stuff.

maybe someone forgot to rebuild these? what iso did you
download?

what exactly did you put in your plan9.ini? our bootloader
does e820 scan for the kernel as the first thing.

you should see something like:

e820="list of big hex numbers here" as the first line printed
by both bootloaders. is that the case?

the bootloader allows you to break into a interactive prompt.

try hitting enter or space just after bios until you see the
">" prompt so you can read the parameters.

the 9pcf kernel on the cdrom and the kernel installed are exactly
the same. you can use the booted kernel from the cdrom to
mount and boot into your installed system by specifying
your install partition on the bootargs prompt.

--
cinap

Re: difference cdrom boot and 9fat boot?

[cinap_lenrek]

the iso is bigger because we include the full hg (mercurial
distributed version control system) repository and because
9front includes python and its dependencies so you can use
hg to update your system.

--
cinap

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

I had exactly the same issue with 9atom (only 8M of memory, then boot
crash while initializing filesystem kfs on 9atom and some zipped fs on
9front when booting from 9fat, but both the live CDs work great). The
difference is that I could rescue the 9atom install by copying the
official plan9 pcf to 9fat. I really have no idea what the difference
really is, but since I had the same issue with both 9atom and 9front,
it should not be 9load vs 9bootfat. The difference between the
official Plan9 pcf (that boots from 9fat) and the 9atom/9front kernels
is that the official one is a bit bigger.

I have tried to set and unset a number of stuff in plan9.ini according
to what I have been able to read up on in
http://plan9.bell-labs.com/wiki/plan9/installation_troubleshooting/diff.html
http://plan9.bell-labs.com/magic/man2html/8/plan9.ini

stuff I have tried to set and unset:
*nomp=1
*maxmem=0xF0000000
*nomce=1
*noe820scan=1
*nopcirouting=1
*nobios=1

My current plan9.ini is quite simple (and a similar one worked with
the Plan9 pcf on 9atom)

bootfile=sdE0!9fat!9pcf
bootargs=local!/dev/sdE0/fscache
sysname=cirno
mouseport=ps2
monitor=vesa
vgasize=1280x800x32

*ncpu=2


About the iso being bigger: The thing I was talking about before was
that the pcf from the official Plan9 iso
(http://plan9.bell-labs.com/plan9/download.html) was bigger than from
9atom or (http://www.quanstro.net/plan9/9atom/) or 9front
(http://aiju.de/9front.iso).

sha1sum for my 9front iso: 26501eb0b21ed07bf2186ebb6b698cf641177223

I am just completely confused about why the cdrom boot would work for
both 9atom and 9front but not 9fat boot...
I am especially eager to get 9front working since that one has network
out-of-the box and I do like the idea of being able to update the
system via Hg :)

Is there documentation somewhere about what the point/reason/aims are
for 9front vs official Plan9 or 9atom? I just stumbled upon this in
search of the broadcom driver :) If I understand it correctly it tries
to modernize the distro and evolve it, which I think looks quite
interesting (especially since I am planning to use it as my hobbyist
OS while learning).

2011/5/31  <cinap_lenrek@gmx.de>:
> the iso is bigger because we include the full hg (mercurial
> distributed version control system) repository and because
> 9front includes python and its dependencies so you can use
> hg to update your system.
>
> --
> cinap
>

Re: difference cdrom boot and 9fat boot?

[cinap_lenrek]

the 9boot bootloader prints all parametrs that it will pass
to the kernel.
so does it print any e820= lines when booting from disk?
 9bootfat?
if not, there are 2 possibilities... 9bootfat was not rebuild
in that iso, or when booting from disk the bios or some side
effect from grub disables the e820 scan.

3) you are not using 9bootfat at all... maybe because the pbs
was not updated and what loads 9pcf is some stale version of
9load that was installed on the fat before.

you can come to irc on freenode.net in the channel #cat-v and
we help you trouble shooting this issue if you want.

--
cinap

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

It does print e820 parameters when booting from disk: lots of zeroes
(I had to boot several times to check this... passes by very fast) and
ends with 9f800

If it is 9bootfat I must have had extremely bad luck since also 9atom
had exactly the same issue.
If it is Grub (which would not surprise me...), it is strange that the
official Plan9 9pcf could boot...

I am starting to wonder if I should aim for a native install of the
official Plan9, then pull the broadcom driver from the 9front CD and
update from there.... Probably this would be faster than figuring this
out :)

About IRC: I will try as soon as I get home from work (IRC blocked
from the work network)

Thanks for all the help!

Cheers
Jens

2011/5/31  <cinap_lenrek@gmx.de>:
> the 9boot bootloader prints all parametrs that it will pass
> to the kernel.
> so does it print any e820= lines when booting from disk?
>  9bootfat?
> if not, there are 2 possibilities... 9bootfat was not rebuild
> in that iso, or when booting from disk the bios or some side
> effect from grub disables the e820 scan.
>
> 3) you are not using 9bootfat at all... maybe because the pbs
> was not updated and what loads 9pcf is some stale version of
> 9load that was installed on the fat before.
>
> you can come to irc on freenode.net in the channel #cat-v and
> we help you trouble shooting this issue if you want.
>
> --
> cinap
>

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

By the way: I also made a note of the stuff printed just before it
stops booting:

8M memory: 6M kernel data 1M user 39M swap
6: bzfs killed: out of memory
boot: can't exec /bin/rc '/bin/rc' does not exist


2011/5/31 Jens Staal <staal1978@gmail.com>:
> It does print e820 parameters when booting from disk: lots of zeroes
> (I had to boot several times to check this... passes by very fast) and
> ends with 9f800
>
> If it is 9bootfat I must have had extremely bad luck since also 9atom
> had exactly the same issue.
> If it is Grub (which would not surprise me...), it is strange that the
> official Plan9 9pcf could boot...
>
> I am starting to wonder if I should aim for a native install of the
> official Plan9, then pull the broadcom driver from the 9front CD and
> update from there.... Probably this would be faster than figuring this
> out :)
>
> About IRC: I will try as soon as I get home from work (IRC blocked
> from the work network)
>
> Thanks for all the help!
>
> Cheers
> Jens
>
> 2011/5/31  <cinap_lenrek@gmx.de>:
>> the 9boot bootloader prints all parametrs that it will pass
>> to the kernel.
>> so does it print any e820= lines when booting from disk?
>>  9bootfat?
>> if not, there are 2 possibilities... 9bootfat was not rebuild
>> in that iso, or when booting from disk the bios or some side
>> effect from grub disables the e820 scan.
>>
>> 3) you are not using 9bootfat at all... maybe because the pbs
>> was not updated and what loads 9pcf is some stale version of
>> 9load that was installed on the fat before.
>>
>> you can come to irc on freenode.net in the channel #cat-v and
>> we help you trouble shooting this issue if you want.
>>
>> --
>> cinap
>>
>

Re: difference cdrom boot and 9fat boot?

[cinap_lenrek]

right, but thats obvious! what is the bootloader printing????

--
cinap

Re: difference cdrom boot and 9fat boot?

[Julius Schmidt]

> I am starting to wonder if I should aim for a native install of the
> official Plan9, then pull the broadcom driver from the 9front CD and
> update from there.... Probably this would be faster than figuring this
> out :)

Yeah but we wouldn't learn anything from it.

> About IRC: I will try as soon as I get home from work (IRC blocked
> from the work network)
So set up some tunnel ;P

aiju

Re: difference cdrom boot and 9fat boot?

[cinap_lenrek]

but it would be more interesting figuring out why this
happens :-)

sure, you could backport bcm driver to bell labs plan9,
but this memory detection problem might happen to others
too... 

one tip (i think i wrote this in my first mail) is that
you can break 9boot into a prompt where it will sit
there until you type boot. just hit any key like space
or enter. (it waits for one second and then does
automatically boot if it had a bootfile= in plan9.ini)

you could also comment out the bootfile= line and
type it at the prompt every time when experimenting.

the command syntax of the prompt is the same
syntax used in plan9.ini (sections/menus are not
supported by 9boot)

it would be very interesting what e820= is and also
the e820 prints from the bell-labs kernel.

i think we have some problem with the e820 call... but
its hard to tell what exactly is wrong without any
data.

if you are interested, the e820= format is just a list
of 2 64bit hex numbers. the first is the start address
of a memory block, and the 2nd is the end address of
the block...

e820= should really be memory= but we had no choice
as 9atom is already using it.

--
cinap

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

[-- Attachment #1: Type: text/plain, Size: 2184 bytes --]

I really start believing that what I got is a GRUB2 bug
Similar issues was seen in this bug report:
http://goo.gl/7j5ll

I did the experiment you suggested to check what the e820 output is.
I wanted to do this as comparable as possible so I have compared
9front booted from CD and booted from 9fat

9fat:
"e820: 0000 0000 0009 f800 memory"

cdboot:
"e820: 0000000000000000 000000000009f800 0000000000100000 000000007fed0000"

I guess the first thing I should try is to boot using another
bootloader (or google for "e820 hooked by grub2" and see if I can
disable it).

Thanks a lot for the help I would never have figured this one out myself!
Now the big mystery is just how the 9pcf from Bell labs plan9 was able
to boot...
I attach the diff of the pcf configuration files for my version of
9pcf from 9front and Bell labs Plan9.
Are there any stuff in there that could influence an overide in RAM
detection in the stock kernel?

2011/5/31  <cinap_lenrek@gmx.de>:
> but it would be more interesting figuring out why this
> happens :-)
>
> sure, you could backport bcm driver to bell labs plan9,
> but this memory detection problem might happen to others
> too...
>
> one tip (i think i wrote this in my first mail) is that
> you can break 9boot into a prompt where it will sit
> there until you type boot. just hit any key like space
> or enter. (it waits for one second and then does
> automatically boot if it had a bootfile= in plan9.ini)
>
> you could also comment out the bootfile= line and
> type it at the prompt every time when experimenting.
>
> the command syntax of the prompt is the same
> syntax used in plan9.ini (sections/menus are not
> supported by 9boot)
>
> it would be very interesting what e820= is and also
> the e820 prints from the bell-labs kernel.
>
> i think we have some problem with the e820 call... but
> its hard to tell what exactly is wrong without any
> data.
>
> if you are interested, the e820= format is just a list
> of 2 64bit hex numbers. the first is the start address
> of a memory block, and the 2nd is the end address of
> the block...
>
> e820= should really be memory= but we had no choice
> as 9atom is already using it.
>
> --
> cinap
>

[-- Attachment #2: pcfdiff --]
[-- Type: application/octet-stream, Size: 437 bytes --]

1c1
< # pcf - pc terminal with fossil root and maybe venti block store
---
> # pcf - pc terminal with local disk
24a25
> 	kbd
26,27d26
< 	kbmap
< 	kbin
56a56
> 	etherbcm        pci
75a76,78
> 	audiosb16	dma
> 	audioac97	audioac97mix
> 
133c136,137
< 	/386/bin/ip/ipconfig
---
> 	/386/bin/bzfs
> 	/386/bin/mntgen
135,138c139
< 	/386/bin/fossil/fossil
< 	/386/bin/venti/venti
< 	/386/bin/usb/usbd
< 	/386/bin/disk/partfs
---
> 	rootfs.bz2

Re: difference cdrom boot and 9fat boot?

[cinap_lenrek]

this is not the output! it cant be...
9bootfat prints in exactly the same form as 9bootiso... it never
prints something like "e820:" and the hex values are alawys
16 hex digits. its the same code after all, the 9bootfat/9bootiso
and 9bootpxe differ only in ther start() functions!

what is this?

cdboot:
"e820: 0000000000000000 000000000009f800 0000000000100000 000000007fed0000"

the numbers there at least look right, but it should really be:

"e820=0000000000000000 000000000009f800 0000000000100000 000000007fed0000"

this one completly puzzles me... 16bit formating? wtf?

9fat:
"e820: 0000 0000 0009 f800 memory"

the "memory" there indicates that this is printed by the kernel i guess?

but what i wanted for like 3 mails is the output of 9bootfats e820 scan!
what does it print for e832=? you said it produces e820= output? what is
it?

did you got to the > prompt with 9bootfat?

this all makes no sense!

:-(

--
cinap

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

Sorry about messing up. I did indeed block the boot by entering the
prompt for each boot option and these were the things I could see....

2011/5/31  <cinap_lenrek@gmx.de>:
> this is not the output! it cant be...
> 9bootfat prints in exactly the same form as 9bootiso... it never
> prints something like "e820:" and the hex values are alawys
> 16 hex digits. its the same code after all, the 9bootfat/9bootiso
> and 9bootpxe differ only in ther start() functions!
>
> what is this?
>
> cdboot:
> "e820: 0000000000000000 000000000009f800 0000000000100000 000000007fed0000"
>
> the numbers there at least look right, but it should really be:
>
> "e820=0000000000000000 000000000009f800 0000000000100000 000000007fed0000"
>
> this one completly puzzles me... 16bit formating? wtf?
>
> 9fat:
> "e820: 0000 0000 0009 f800 memory"
>
> the "memory" there indicates that this is printed by the kernel i guess?
>
> but what i wanted for like 3 mails is the output of 9bootfats e820 scan!
> what does it print for e832=? you said it produces e820= output? what is
> it?
>
> did you got to the > prompt with 9bootfat?
>
> this all makes no sense!
>
> :-(
>
> --
> cinap
>

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

I checked again and it is indeed as you said.
From the cdboot you got e820= (numbers) as the first line visible on screen
from the 9fat boot it is only the e820: (numbers) after a couple of
other lines (so probably kernel output).

I guess it may then be something weird with my 9bootfat. I will pull
another iso and copy that 9bootfat over and see what happens.

I also (naively) just tried to define
e820=(numbers)
in plan9.ini circumvent whatever is wrong, but that did not work. Are
you aware of some kind of similar option?

2011/6/1 Jens Staal <staal1978@gmail.com>:
> Sorry about messing up. I did indeed block the boot by entering the
> prompt for each boot option and these were the things I could see....
>
> 2011/5/31  <cinap_lenrek@gmx.de>:
>> this is not the output! it cant be...
>> 9bootfat prints in exactly the same form as 9bootiso... it never
>> prints something like "e820:" and the hex values are alawys
>> 16 hex digits. its the same code after all, the 9bootfat/9bootiso
>> and 9bootpxe differ only in ther start() functions!
>>
>> what is this?
>>
>> cdboot:
>> "e820: 0000000000000000 000000000009f800 0000000000100000 000000007fed0000"
>>
>> the numbers there at least look right, but it should really be:
>>
>> "e820=0000000000000000 000000000009f800 0000000000100000 000000007fed0000"
>>
>> this one completly puzzles me... 16bit formating? wtf?
>>
>> 9fat:
>> "e820: 0000 0000 0009 f800 memory"
>>
>> the "memory" there indicates that this is printed by the kernel i guess?
>>
>> but what i wanted for like 3 mails is the output of 9bootfats e820 scan!
>> what does it print for e832=? you said it produces e820= output? what is
>> it?
>>
>> did you got to the > prompt with 9bootfat?
>>
>> this all makes no sense!
>>
>> :-(
>>
>> --
>> cinap
>>
>

Re: difference cdrom boot and 9fat boot?

[Jens Staal]

SOLVED!

This is both embarrasing and a bit surprising, but I think your
hypothesis of "contamination" from previous plan9 installation at the
same partition somehow was right.

When I tried reinstalling before I usually just deleted all the plan9
partitions in the installation. Now I removed the partition in Linux,
formatted the empty space in a new partition and deleted that one too.

After installing on this patch of empty space, the 9Front boots just
fine. I am extremely surprised and none the wiser.

On the bright side: now I might be able to play with stuff that I at
least can hope to begin to understand (or at least where I can get
hold of the errors ;) )

2011/6/1 Jens Staal <staal1978@gmail.com>:
> I checked again and it is indeed as you said.
> From the cdboot you got e820= (numbers) as the first line visible on screen
> from the 9fat boot it is only the e820: (numbers) after a couple of
> other lines (so probably kernel output).
>
> I guess it may then be something weird with my 9bootfat. I will pull
> another iso and copy that 9bootfat over and see what happens.
>
> I also (naively) just tried to define
> e820=(numbers)
> in plan9.ini circumvent whatever is wrong, but that did not work. Are
> you aware of some kind of similar option?
>
> 2011/6/1 Jens Staal <staal1978@gmail.com>:
>> Sorry about messing up. I did indeed block the boot by entering the
>> prompt for each boot option and these were the things I could see....
>>
>> 2011/5/31  <cinap_lenrek@gmx.de>:
>>> this is not the output! it cant be...
>>> 9bootfat prints in exactly the same form as 9bootiso... it never
>>> prints something like "e820:" and the hex values are alawys
>>> 16 hex digits. its the same code after all, the 9bootfat/9bootiso
>>> and 9bootpxe differ only in ther start() functions!
>>>
>>> what is this?
>>>
>>> cdboot:
>>> "e820: 0000000000000000 000000000009f800 0000000000100000 000000007fed0000"
>>>
>>> the numbers there at least look right, but it should really be:
>>>
>>> "e820=0000000000000000 000000000009f800 0000000000100000 000000007fed0000"
>>>
>>> this one completly puzzles me... 16bit formating? wtf?
>>>
>>> 9fat:
>>> "e820: 0000 0000 0009 f800 memory"
>>>
>>> the "memory" there indicates that this is printed by the kernel i guess?
>>>
>>> but what i wanted for like 3 mails is the output of 9bootfats e820 scan!
>>> what does it print for e832=? you said it produces e820= output? what is
>>> it?
>>>
>>> did you got to the > prompt with 9bootfat?
>>>
>>> this all makes no sense!
>>>
>>> :-(
>>>
>>> --
>>> cinap
>>>
>>
>

Re: difference cdrom boot and 9fat boot?

[Uriel]

Just a few useless (and unusually happy) comments.

On Tue, May 31, 2011 at 11:39 AM,  <cinap_lenrek@gmx.de> wrote:
> this is really wired... the iso and fat bootloaders (we
> dont use 9load) share most of the code except filesystem
> specific stuff.
>
> maybe someone forgot to rebuild these? what iso did you
> download?

We need a more clear ISO-blessing process, yes, it is 'bureaucracy',
but people need to know what the fuck they are downloading and which
are the latest official bits.

> the bootloader allows you to break into a interactive prompt.
>
> try hitting enter or space just after bios until you see the
> ">" prompt so you can read the parameters.

I knew this, but still makes me so happy that at last it is done.

> the 9pcf kernel on the cdrom and the kernel installed are exactly
> the same. you can use the booted kernel from the cdrom to
> mount and boot into your installed system by specifying
> your install partition on the bootargs prompt.

And *fuck yea* for this, that BL never managed to do this is a
disgrace, yay for one kernel to rule them all!

uriel