[9fans] Cannot boot plan9

[9fans] Cannot boot plan9

[Szymon Olewniczak]

[-- Attachment #1: Type: text/plain, Size: 337 bytes --]

Hi,
today I've tried to boot plan9 on my computer. Sadly withou a success.
I'm attaching screenshot of my computer after typing ctrl+t ctrl+t p.

Do you know is going on? How can I fix it? I woud like to run Plan 9
before 21.03 in order to see how it's working and put an application to some
of the GSoC project :)

BR,
Szymon

[-- Attachment #2: crash.png --]
[-- Type: image/png, Size: 6189951 bytes --]

Re: [9fans] Cannot boot plan9

[erik quanstrom]

On Sun Mar 16 12:06:39 EDT 2014, szymon.olewniczak@rid.pl wrote:

> Hi,
> today I've tried to boot plan9 on my computer. Sadly withou a success.
> I'm attaching screenshot of my computer after typing ctrl+t ctrl+t p.
>
> Do you know is going on? How can I fix it? I woud like to run Plan 9
> before 21.03 in order to see how it's working and put an application to some
> of the GSoC project :)

well, it looks like it can't find anything in #S, which is the generic disk
driver.  this means your hardware wasn't recognized.  if you can boot
the 9atom usb installer, even if you can't see the disks, maybe we can
see what's wrong.  i'd like to see the output of "pci|grep disk".

the usb installer is here

	http://ftp.9atom.org/other/+usbinstamd64.bz2

just bunzip2 the image, and dd the result to a usb stick that's at least
512MB.

this information should give enough information to yield a patch
to the distribution, and will satisfy gsoc by itself.

- erik

Re: [9fans] Cannot boot plan9

[Szymon Olewniczak]

On Sun, Mar 16, 2014 at 12:14:17PM -0400, erik quanstrom wrote:
> On Sun Mar 16 12:06:39 EDT 2014, szymon.olewniczak@rid.pl wrote:
>
> > Hi,
> > today I've tried to boot plan9 on my computer. Sadly withou a success.
> > I'm attaching screenshot of my computer after typing ctrl+t ctrl+t p.
> >
> > Do you know is going on? How can I fix it? I woud like to run Plan 9
> > before 21.03 in order to see how it's working and put an application to some
> > of the GSoC project :)
>
> well, it looks like it can't find anything in #S, which is the generic disk
> driver.  this means your hardware wasn't recognized.  if you can boot
> the 9atom usb installer, even if you can't see the disks, maybe we can
> see what's wrong.  i'd like to see the output of "pci|grep disk".
Here it is:

0.6.0 disk 01.01.8a 10de/03ec 0 4:0000ffa1 16
0.8.0 disk 01.01.85 10de/03f6 5 0:0000d401 16 1:0000d081 16 2:0000d001
16 3:0000cc01 16 4:0000c881 16 5:deefc000 4096

But I have no idea what does it means. I hope you will know.

> the usb installer is here
>
> 	http://ftp.9atom.org/other/+usbinstamd64.bz2
>
> just bunzip2 the image, and dd the result to a usb stick that's at least
> 512MB.
Wow! It's working! What is the diffirence between standard CD version and this one?

> this information should give enough information to yield a patch
> to the distribution, and will satisfy gsoc by itself.
Do you meen that I can create a driver to my hd as the gosc project?

BR,
Szymon

Re: [9fans] Cannot boot plan9

[erik quanstrom]

> 0.6.0 disk 01.01.8a 10de/03ec 0 4:0000ffa1 16 0.8.0 disk 01.01.85
> 10de/03f6 5 0:0000d401 16 1:0000d081 16 2:0000d001 16 3:0000cc01 16
> 4:0000c881 16 5:deefc000 4096
>
> But I have no idea what does it means.  I hope you will know.

i do.  this is an nvidia sata part, which is not currently supported
by plan 9.

> > the usb installer is here
> >
> > 	http://ftp.9atom.org/other/+usbinstamd64.bz2
> >
> > just bunzip2 the image, and dd the result to a usb stick that's at
> > least 512MB.
> Wow!  It's working!  What is the diffirence between standard CD
> version and this one?

the difference is that in not using a cdrom, we avoid the issue of
a missing driver for the sata_nv part.

> > this information should give enough information to yield a patch to
> > the distribution, and will satisfy gsoc by itself.
> Do you meen that I can create a driver to my hd as the gosc project?

well, you could propose that, but i don't think you could get a mentor
really excited about that project since it is (a) uncommon (b) last produced
many years ago (c) undocumented outside linux drivers.

linux sata drivers can be harder to clone since so much of the work is done
in a series of generic layers.  and they can take a while to digest.

i would recommend running from usb on this machine.  alternatively,
find an ahci add in card.

- erik

Re: [9fans] Cannot boot plan9

[cinap_lenrek]

this device is a ata controller that might just work with
plan9's generic ata driver (given you add the pci id).
i checked 9front sdide.c which has the pci id for the controller:

                case (0x03ec<<16)|0x10DE:       /* nVidia nForce 61 MCP SATA */

we add these pci id's as we encounter the device so it has been
tested and worked (at least at some point) with 9front.

so i highly recommend just adding the pci device id to your ide/ata
driver and try it out. this is not some rocket science justifying a
gsoc project. :)

--
cinap

Re: [9fans] Cannot boot plan9

[erik quanstrom]

On Tue Mar 18 06:53:05 EDT 2014, cinap_lenrek@felloff.net wrote:
> this device is a ata controller that might just work with
> plan9's generic ata driver (given you add the pci id).
> i checked 9front sdide.c which has the pci id for the controller:
>
>                 case (0x03ec<<16)|0x10DE:       /* nVidia nForce 61 MCP SATA */
>
> we add these pci id's as we encounter the device so it has been
> tested and worked (at least at some point) with 9front.

you had me worried that i'd slipped a cog.

0x03ec is in sources kernel.  i sent in a patch in 2007.  it was
added aug 5, 2007.  it's in 9front because it was in the code
9front started with.  unless i've missed something, adding this
vid/did isn't going to help.

- erik

---
; cat email
; pwd
/n/sources/patch/applied/sdata-mcp6x
; ls -ld .
d-rwxrwxr-x M 325 geoff sys 0 Aug  5  2007 sdata-mcp6x
; cat email
quanstro@quanstro.net
; diff -c sdata.c.orig sdata.c.new
sdata.c.orig:1970,1976 - sdata.c.new:1970,1979
  		case (0x0054<<16)|0x10DE:	/* nVidia nForce4 SATA */
  		case (0x0055<<16)|0x10DE:	/* nVidia nForce4 SATA */
  		case (0x0266<<16)|0x10DE:	/* nVidia nForce4 430 SATA */
- 		case (0x0267<<16)|0x10DE:	/* nVidia nForce4 430 SATA */
+ 		case (0x0267<<16)|0x10DE:	/* nVidia nForce 55 MCP SATA */
+ 		case (0x03EC<<16)|0x10DE:	/* nVidia nForce 61 MCP SATA */
+ 		case (0x0448<<16)|0x10DE:	/* nVidia nForce 65 MCP SATA */
+ 		case (0x0560<<16)|0x10DE:	/* nVidia nForce 69 MCP SATA */
  			/*
  			 * Ditto, although it may have a different base
  			 * address for the registers (0x50?).


; history -d /n/sources/plan9/sys/src/9/pc/sdata.c
	[...]
Aug  5 20:52:05 EDT 2007 /n/sourcesdump/2008/0205/plan9/sys/src/9/pc/sdata.c 52991 [geoff]
/n/sourcesdump/2007/0805/plan9/sys/src/9/pc/sdata.c:1970,1976 - /n/sourcesdump/2008/0205/plan9/sys/src/9/pc/sdata.c:1970,1979
  		case (0x0054<<16)|0x10DE:	/* nVidia nForce4 SATA */
  		case (0x0055<<16)|0x10DE:	/* nVidia nForce4 SATA */
  		case (0x0266<<16)|0x10DE:	/* nVidia nForce4 430 SATA */
- 		case (0x0267<<16)|0x10DE:	/* nVidia nForce4 430 SATA */
+ 		case (0x0267<<16)|0x10DE:	/* nVidia nForce 55 MCP SATA */
+ 		case (0x03EC<<16)|0x10DE:	/* nVidia nForce 61 MCP SATA */
+ 		case (0x0448<<16)|0x10DE:	/* nVidia nForce 65 MCP SATA */
+ 		case (0x0560<<16)|0x10DE:	/* nVidia nForce 69 MCP SATA */
  			/*
  			 * Ditto, although it may have a different base
  			 * address for the registers (0x50?).

Re: [9fans] Cannot boot plan9

[cinap_lenrek]

yep. sorry, wrong assumptions on my side. i only
have reports of MCP51 working, but not MCP61.

--
cinap

Re: [9fans] Cannot boot plan9

[erik quanstrom]

On Tue Mar 18 10:33:08 EDT 2014, cinap_lenrek@felloff.net wrote:
> yep. sorry, wrong assumptions on my side. i only
> have reports of MCP51 working, but not MCP61.

i believe both work since iirc, i had a mcp61 motherboard,
and used the ide interface.

iirc, nvidia is not like intel.  the ide interface was hardwired to
a pata port.  the nv_sata interface was hardwired to a
sata port.

- erik

Re: [9fans] Cannot boot plan9

[Szymon Olewniczak]

> > 0.6.0 disk 01.01.8a 10de/03ec 0 4:0000ffa1 16 0.8.0 disk 01.01.85
> > 10de/03f6 5 0:0000d401 16 1:0000d081 16 2:0000d001 16 3:0000cc01 16
> > 4:0000c881 16 5:deefc000 4096
> >
> > But I have no idea what does it means.  I hope you will know.
>
> i do.  this is an nvidia sata part, which is not currently supported
> by plan 9.
>
> > > the usb installer is here
> > >
> > > 	http://ftp.9atom.org/other/+usbinstamd64.bz2
> > >
> > > just bunzip2 the image, and dd the result to a usb stick that's at
> > > least 512MB.
> > Wow!  It's working!  What is the diffirence between standard CD
> > version and this one?
>
> the difference is that in not using a cdrom, we avoid the issue of
> a missing driver for the sata_nv part.
>
> > > this information should give enough information to yield a patch to
> > > the distribution, and will satisfy gsoc by itself.
> > Do you meen that I can create a driver to my hd as the gosc project?
>
> well, you could propose that, but i don't think you could get a mentor
> really excited about that project since it is (a) uncommon (b) last produced
> many years ago (c) undocumented outside linux drivers.
>
> linux sata drivers can be harder to clone since so much of the work is done
> in a series of generic layers.  and they can take a while to digest.
Do you think that there is a point in trying to solve it? Where should I
start when I would like to fix this issue. Is it easy or rather hard
task? How does the develop process looks like when doing such things?

> i would recommend running from usb on this machine.
Does plan9 suport something like linux tmpfs? because when I boot
the system from the image you have given me I cannot save any file. Does
plan9 support saving the state of live session?

> alternatively, find an ahci add in card.
It's probably a noob question but how can I do this? Is it's controled
by some bios software or it needs some hardware changes?

Re: [9fans] Cannot boot plan9

[erik quanstrom]

> Do you think that there is a point in trying to solve it? Where should I
> start when I would like to fix this issue. Is it easy or rather hard
> task? How does the develop process looks like when doing such things?

i think this is a hard question to answer.  on the one hand, i don't want
to be discouraging or anything like that.  but on the other hand, it's a
little bit like a mensa question.  it's possible to get really stuck.  and there's
the h/w availability issue.

if you really want to do it, why don't you write a proposal.

since there are no docs, you will have to get the linux or *bsd driver,
and start disecting it.
1.  understand the general scope
2.  find bits like toe that can be completely dropped.
3.  work out what the os interface is.

since this gets boring and i tend to get antsy without working code,
i generally start writing a skeleton driver.  think of wirth's example of
writing a program.  it's a pretty good approach for drivers.

i'm sure you can recognize (but not drive) the hardware on the first day.
break things down into steps so that each bit of code adds functionality.
resist at all costs the temptation to have nothing working till the last
semicolon is in place.

that's my recommendation.

> > i would recommend running from usb on this machine.
> Does plan9 suport something like linux tmpfs?

ramfs.  on plan 9 this is a user-level process. you can run it at any time,
and concurrently run several in different namespaces.

> because when I boot
> the system from the image you have given me I cannot save any file. Does
> plan9 support saving the state of live session?

that's a bit surprising.

try
	echo >/lib/emptyfile

does that yield an error?

> > alternatively, find an ahci add in card.
> It's probably a noob question but how can I do this? Is it's controled
> by some bios software or it needs some hardware changes?

each add-in card determines the register interface it would like to present.
bios has no say in the matter.

- erik

Re: [9fans] Cannot boot plan9

[Szymon Olewniczak]

> > because when I boot
> > the system from the image you have given me I cannot save any file. Does
> > plan9 support saving the state of live session?
>
> that's a bit surprising.
>
> try
> 	echo >/lib/emptyfile
>
> does that yield an error?
Ok, now it's working. I'm not sure why it doesn't work last time(maybe I
was doing something wrong). But now I have several more questions about
this image:
At first it seems that not every plan 9 programs are avalible. For
example man doesn't work becouse it cannot run /bin/troff. Also the
"page" program seems not to be avalible here.

Second question: how can I mount pendrive in Plan 9. I've tried to
follow steps provided by [1] but it seems that there aren't any binaries in
/bin/usb directory. What is going wrong?

[1] http://man.cat-v.org/plan_9/4/usbdisk

Re: [9fans] Cannot boot plan9

[erik quanstrom]

> Ok, now it's working. I'm not sure why it doesn't work last time(maybe I
> was doing something wrong). But now I have several more questions about
> this image:

easy for little errors to creep in when one is unsure of the details.
(as i'm experiencing right now with something else.)

> At first it seems that not every plan 9 programs are avalible. For
> example man doesn't work becouse it cannot run /bin/troff. Also the
> "page" program seems not to be avalible here.

if you're using the atom usb install image, it is very sparse.  only the programs
required for install are included.  you can compile the rest of the system,
though, since all the source is there.

> Second question: how can I mount pendrive in Plan 9. I've tried to
> follow steps provided by [1] but it seems that there aren't any binaries in
> /bin/usb directory. What is going wrong?
>
> [1] http://man.cat-v.org/plan_9/4/usbdisk

it should mount itself.  do you see usb/disk... in the console window?
what about "cat /dev/kmesg"?

- erik

Re: [9fans] Cannot boot plan9

[Szymon Olewniczak]

[-- Attachment #1: Type: text/plain, Size: 1679 bytes --]

On Fri, Mar 21, 2014 at 06:38:21PM -0400, erik quanstrom wrote:
> > Ok, now it's working. I'm not sure why it doesn't work last time(maybe I
> > was doing something wrong). But now I have several more questions about
> > this image:
>
> easy for little errors to creep in when one is unsure of the details.
> (as i'm experiencing right now with something else.)
>
> > At first it seems that not every plan 9 programs are avalible. For
> > example man doesn't work becouse it cannot run /bin/troff. Also the
> > "page" program seems not to be avalible here.
>
> if you're using the atom usb install image, it is very sparse.  only the programs
> required for install are included.  you can compile the rest of the system,
> though, since all the source is there.
>
> > Second question: how can I mount pendrive in Plan 9. I've tried to
> > follow steps provided by [1] but it seems that there aren't any binaries in
> > /bin/usb directory. What is going wrong?
> >
> > [1] http://man.cat-v.org/plan_9/4/usbdisk
>
> it should mount itself.  do you see usb/disk... in the console window?

usb/disk: '/bin/usb/disk' does not exist

In addition I cannot boot plan9 atom when there are two usb disks connected to my PC.
boot(8) just reach the point when it promt for root and when I press enter the computer
reboots. Do you know how to solve it?

> what about "cat /dev/kmesg"?
I've attached image of the result of the command. I hope that soon I will be able to
attach it in plain text format :). I'm afraid that my Ethernet card is not recognised too.
Can you recommend any PCI Ethernet card that is known to be working under Plan 9?

BR,
Szymon

[-- Attachment #2: IMG_2193.JPG --]
[-- Type: image/jpeg, Size: 1091015 bytes --]

[-- Attachment #3: IMG_2194.JPG --]
[-- Type: image/jpeg, Size: 1450733 bytes --]

Re: [9fans] Cannot boot plan9

[erik quanstrom]

> I've attached image of the result of the command. I hope that soon I will be able to
> attach it in plain text format :). I'm afraid that my Ethernet card is not recognised too.
> Can you recommend any PCI Ethernet card that is known to be working under Plan 9?

anything intel gigabit+pcie.  most 8169 (realtek gigabit).

- erik

Re: [9fans] Cannot boot plan9

[Alex Jordan]

[-- Attachment #1: Type: text/plain, Size: 447 bytes --]

On Mar 17, 2014 3:34 PM, "erik quanstrom" <quanstro@quanstro.net> wrote:
> linux sata drivers can be harder to clone since so much of the work is
done
> in a series of generic layers.  and they can take a while to digest.
As an aside, why don't we have an architecture like this? (From your
comment, I'm assuming we don't.)
I'm sure there are valid technical reasons, but to an uneducated bystander
like myself, this seems to violate DRY.

[-- Attachment #2: Type: text/html, Size: 558 bytes --]

Re: [9fans] Cannot boot plan9

[erik quanstrom]

> As an aside, why don't we have an architecture like this? (From your
> comment, I'm assuming we don't.)
> I'm sure there are valid technical reasons, but to an uneducated bystander
> like myself, this seems to violate DRY.

because at least in the case of the linux disk layers, it makes the code harder
to read, follow and maintain.  code is read many times, and written once.
therefore, readability is more important than writability.  plan 9 tends to
avoid middle-management code.

heavily layered code can also be hard to test.

certainly, repeated code isn't something one wishes to create.  and plan 9
has too much of it.  on the other hand, combining code isn't free.

one of the ideas that go has had is to use interfaces extensively.
this allows fine-grained reuse, without the worry that management code has.

if you've worked for even a moderately large orginization, perhaps you can
relate to the difficulties of middle-managment code.  :-)

- erik