Toshiba AC100
Phil Endecott
As many of you will know the Toshiba AC100 is a netbook based on an
Nvidia Tegra ARM processor that ships with Android. I've been waiting
a long time for hardware like this to become available; I've just
bought one, and it is excellent. It's light, cool, and great quality.
On the other hand, the shipped software is not going to please anyone;
the decision to go with Android just seems bizarre.
Of course it didn't take long for someone to hack it. It seems that
it's sufficiently close to Nvidia's reference design that it's quite
easy to boot an alternative OS. The best resource at present seems to
be here: http://ac100.gudinna.com/README . As with a lot of projects
of this sort, the people involved seem to be rather spread out and can
be found at Toshiba's own forums
(http://forums.computers.toshiba-europe.com/forums/) (which I would
avoid because posts seem to get edited by Toshiba if you say the wrong
thing) and at this site: http://tosh-ac100.wetpaint.com/ (which I've
not registered with because it wanted to know my age and gender!). The
author of the gudinna.com page doesn't seem to want to identify
him/herself for some reason.
The procedure is basically as follows: you get an Nvidia x86 binary,
whose license says that it can only be used with their development
hardware, and use it over USB to replace one partition of the internal
flash. After this modification the device will scan the SD card during
boot. So this is a fairly non-intrusive change that lets you continue
to boot Android. However, the SD card is (presumably) slower than the
internal flash. You then install your OS of choice on the SD card and
off you go. An Ubuntu image is available from
http://ac100.gudinna.com/ (you need to use --numeric-owner when
untaring on a Debian system else it will fail to work in some nasty
subtle way). This naturally has some wrinkles at the moment but I
guess they will be smoothed out soon.
Sooo..... who is interested in getting Debian to work on this
machine? I guess there are a few issues:
1. Specific to this machine, there are issues related to installation.
Those are being resolved already by others, so we can copy that. One
potential issue is the legal one with the Nvidia tool; it may be that
it's simple to reverse-engineer the protocol it uses over USB, or something.
2. Presumably there will also be machine-specific driver issues, but
most things seem to be working in Ubuntu so that might not be too
hard. I'm not sure how many binary blobs are involved.
3. Then there's the whole ARM architecture version thing. Each time I
look at this I realise it's more complicated that I had hoped, for
example I have read that the Tegra processor in this machine doesn't
have NEON. Debian obviously doesn't want to have 42 different ARM
architecture variants. My guess is that in the case of the FP variants
there are only relatively few packages that are really fp-intensive, so
a scheme that allows them to have optimised versions distributed
separately might be best, and have the core architecture just cover the
"ARMv7" aspect.
I hope there are others out there interested in this machine, the first
of its kind.
Cheers, Phil.
(BTW I'm in Cambridge, and if anyone would like to play with it you're welcome.)
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/128656...@dmwebmail.dmwebmail.chezphil.org
Wookey
> Hi Everyone,
>
> As many of you will know the Toshiba AC100 is a netbook based on an
> Nvidia Tegra ARM processor that ships with Android. I've been waiting a
> long time for hardware like this to become available; I've just bought
> one, and it is excellent. It's light, cool, and great quality. On the
> other hand, the shipped software is not going to please anyone; the
> decision to go with Android just seems bizarre.
I saw one today too, as Peter Maydell's has just arrived. He's got
ubuntu Maverick going on it, booting from SD, and apparently running
the Adnroid kernel. Seems to work fine in general, but no wifi yet so
limited experimentation.
> Sooo..... who is interested in getting Debian to work on this machine?
> I guess there are a few issues:
> 3. Then there's the whole ARM architecture version thing. Each time I
> look at this I realise it's more complicated that I had hoped, for
> example I have read that the Tegra processor in this machine doesn't
> have NEON.
That's correct.
> Debian obviously doesn't want to have 42 different ARM
> architecture variants.
No, we don't.
> My guess is that in the case of the FP variants
> there are only relatively few packages that are really fp-intensive, so
> a scheme that allows them to have optimised versions distributed
> separately might be best, and have the core architecture just cover the
> "ARMv7" aspect.
Absolutely - and this is on my list of things to get linaro/ubuntu to
spend some time on, although I've been told nothing much is going to
happen on that front until multiarch is actually done, so it'll be a
while yet before there is much in the way of practical improvements.
It should start to become clear over the next few months how many
variants are actually worth supporting. That'll be something at least.
> I hope there are others out there interested in this machine, the first
> of its kind.
(apart from the pegatron machines and the Genesi Smartbook:
http://www.genesi-usa.com/press/2010/8/27/ )
> (BTW I'm in Cambridge, and if anyone would like to play with it you're welcome.)
So that's at least 2 in Cambridge :-)
Wookey
--
Principal hats: Linaro, Emdebian, Wookware, Balloonboard, ARM
http://wookware.org/
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/2010100821...@dream.aleph1.co.uk
Konstantinos Margaritis
> > My guess is that in the case of the FP variants
> > there are only relatively few packages that are really fp-intensive, so
> > a scheme that allows them to have optimised versions distributed
> > separately might be best, and have the core architecture just cover the
> > "ARMv7" aspect.
>
> Absolutely - and this is on my list of things to get linaro/ubuntu to
> spend some time on, although I've been told nothing much is going to
> happen on that front until multiarch is actually done, so it'll be a
> while yet before there is much in the way of practical improvements.
> It should start to become clear over the next few months how many
> variants are actually worth supporting. That'll be something at least.
In the meantime he could try the armhf port on debian-ports.org, which targets
exactly those base requirements: armv7-a, hardfloat ABI (vs softfp), vfpv3-d16
and thumb2. No NEON is required by default so it should run quite fine on the
tegra2. Instead of providing just a bunch of optimized packages, every binary
was rebuilt using the new abi and base specs. Performance has been found
10-40% faster than equivalent softfp code, in some rare cases 200% gain has
been observed.
There is no installer yet but I'll upload an updated tarball on
http://freevec.org/repository/images/
(There are already a couple there, but they're a bit outdated and I intend to
provide a new version during the weekend). I'll post the url here on Sunday.
IMHO, starting a new port is a much faster process than waiting for multiarch
-it took me just a month to build 60% of the debian archive, even with lots of
problems and it keeps on compiling.
My 2c.
Regards
Konstantinos
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/201010090127...@genesi-usa.com
Peter Maydell
> +++ Phil Endecott [2010-10-08 21:24 +0100]:
>> Nvidia Tegra ARM processor that ships with Android. I've been waiting a
>> long time for hardware like this to become available; I've just bought
>> one, and it is excellent.
> I saw one today too, as Peter Maydell's has just arrived. He's got
> ubuntu Maverick going on it, booting from SD, and apparently running
> the Adnroid kernel. Seems to work fine in general, but no wifi yet so
> limited experimentation.
Yep (not that I did any of the hard work, I just installed according to
the information on http://tosh-ac100.wetpaint.com/page/Ubuntu).
Wifi does work, and the machine is basically fine now for the sort
of random spodding (irc, usenet, web browsing) I wanted it for :-)
There's a TODO list on that wiki, of which I think the most important
missing bits are sound and all the power management related stuff.
(At the moment on ubuntu I get about 3 1/2 hours out of the battery
because it never turns anything off. Also lack of 'suspend' is a bit
awkward.) And at some point the boot sequence ought to be smoothed
out a bit so you don't have to use the magic 'recovery mode' nerve
pinch to get it to boot the other OS rather than Android.
-- PMM
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTimYXkJWsvh180_CB...@mail.gmail.com
Wookey
> IMHO, starting a new port is a much faster process than waiting for multiarch
> -it took me just a month to build 60% of the debian archive, even with lots of
> problems and it keeps on compiling.
Of course it is, but this is meaningless comparison; they are not
alternative things, they are complimentary things. A new architecture
(port) is a new architecture. Multiarch is a way of installing things
from more than one arch at a time, and satisfying dependencies between
them.
And the flavour support I was muttering about is something else again
- to make it possible to build, install and use different
optimisations/ISA levels within one architecture, rather than having
to make a whole new arch for every one.
Wookey
--
Principal hats: Linaro, Emdebian, Wookware, Balloonboard, ARM
http://wookware.org/
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/2010100916...@dream.aleph1.co.uk
Luke Kenneth Casson Leighton
by toshiba digital media
group, by CD. i have uploaded them here:
https://alioth.debian.org/scm/?group_id=100475
the git branch is ac100/2.6.29/lkcl
l.
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTikEWMN3vCtsV+kHG...@mail.gmail.com
Maciej Grela
> i have received the Toshiba AC100 GPL kernel sources, kindly supplied
> by toshiba digital media
> group, by CD. i have uploaded them here:
>
> https://alioth.debian.org/scm/?group_id=100475
>
> the git branch is ac100/2.6.29/lkcl
>
Cannot clone this repo:
grela@kraken ~ $ git clone
https://alioth.debian.org/anonscm/git/arm-netbook/arm-netbook.git
Cloning into arm-netbook...
warning: remote HEAD refers to nonexistent ref, unable to checkout.
grela@kraken ~ $
--
Maciej Grela
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTiniGqsQ0KOL_z4Eo...@mail.gmail.com
Luke Kenneth Casson Leighton
> 2010/10/18 Luke Kenneth Casson Leighton <lk...@lkcl.net>:
>> i have received the Toshiba AC100 GPL kernel sources, kindly supplied
>> by toshiba digital media
>> group, by CD. i have uploaded them here:
>>
>> https://alioth.debian.org/scm/?group_id=100475
>>
>> the git branch is ac100/2.6.29/lkcl
>>
>
> Cannot clone this repo:
>
> grela@kraken ~ $ git clone
> https://alioth.debian.org/anonscm/git/arm-netbook/arm-netbook.git
> Cloning into arm-netbook...
> warning: remote HEAD refers to nonexistent ref, unable to checkout.
>
> grela@kraken ~ $
vvvvvvvvvvvvvvvvvvvvvvvvvv
>> the git branch is ac100/2.6.29/lkcl
^^^^^^^^^^^^^^^^^^^^
l.
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTik-FY2JE+BVipqnv...@mail.gmail.com
Luke Kenneth Casson Leighton
<lk...@lkcl.net> wrote:
> On Tue, Oct 19, 2010 at 9:44 PM, Maciej Grela <maciej...@gmail.com> wrote:
>> 2010/10/18 Luke Kenneth Casson Leighton <lk...@lkcl.net>:
>>> i have received the Toshiba AC100 GPL kernel sources, kindly supplied
>>> by toshiba digital media
>>> group, by CD. i have uploaded them here:
>>>
>>> https://alioth.debian.org/scm/?group_id=100475
>>>
>>> the git branch is ac100/2.6.29/lkcl
>>>
>>
>> Cannot clone this repo:
>>
>> grela@kraken ~ $ git clone
>> https://alioth.debian.org/anonscm/git/arm-netbook/arm-netbook.git
>> Cloning into arm-netbook...
>> warning: remote HEAD refers to nonexistent ref, unable to checkout.
>>
>> grela@kraken ~ $
>
> vvvvvvvvvvvvvvvvvvvvvvvvvv
>>> the git branch is ac100/2.6.29/lkcl
> ^^^^^^^^^^^^^^^^^^^^
maciej,
i haven't received a "thank you" or any kind of acknowledgement from
you, so i do not know if you have understood the above repeated
information. please firstly let me apologise for assuming that you're
sufficiently unintelligent to understand why i repeated that
information, but my concern is genuine in that i'd like you to be able
to gain access to the source code. now let me spell it out for you.
once you have done the git clone command, you have the entire
repository. git then assumes, having downloaded the entire
repository, that there is a "master" branch, which it proceeds with
attempting to check out. the repository however simply does not have
a "master" branch - specifically, it does not have a branch named
"master" because... well.. there _is_ no linux branch named "master",
and that's the end of it.
so, after having done that command, you simply do this second command:
git checkout -b ac100/2.6.29/lkcl
is that clear enough?
l.
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTinfCo3GUNAtx84_b...@mail.gmail.com
Oliver Grawert
Am Freitag, den 08.10.2010, 21:24 +0100 schrieb Phil Endecott:
> I hope there are others out there interested in this machine, the first
> of its kind.
...
great to see that there is debian interest, as one of the devs who work
on the image, i'd like to point you guys to some missing points ...
the ac100 devs usually meet in #ac100 on freenode ...
there is a kernel tree at http://gitorious.org/ac100/
thats mainly teh one people contribute too (we currently have some
gentoo and OE devs as well as ubuntu guys ... debian people would be
most welcome)
there is also a mailing list ac...@lists.launchpad.net hosted under the
launchpad.net/ac100 project ...
ciao
oli
Luke Kenneth Casson Leighton
> hi,
> Am Freitag, den 08.10.2010, 21:24 +0100 schrieb Phil Endecott:
>
>> I hope there are others out there interested in this machine, the first
>> of its kind.
> ...
> great to see that there is debian interest, as one of the devs who work
> on the image, i'd like to point you guys to some missing points ...
>
> the ac100 devs usually meet in #ac100 on freenode ...
> there is a kernel tree at http://gitorious.org/ac100/
fantastic. it would be good to let people know of the existence of
the official kernel sources from toshiba. could you kindly do that?
> thats mainly teh one people contribute too (we currently have some
> gentoo and OE devs as well as ubuntu guys ... debian people would be
> most welcome)
>
> there is also a mailing list ac...@lists.launchpad.net hosted under the
> launchpad.net/ac100 project ...
ok i found its mailing list archives (manually) - they're empty.
reporting a bug is not possible because of "you must have an ac100
kernel configured" bullshit launchpad logic.
the tree appears to have toshiba ac100 patches but on closer
inspection it appears instead, at a cursory glance, to be nv-patches
not ac100 patches.
so... i believe that the project could benefit greatly from knowing
of the existence of the official gpl linux kernel code released
officially by toshiba, but it is somewhat ... challenging shall we say
to inform them of that fact.
l.
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTi=H49et+=5ss1x=y5bAJg8+iQuF...@mail.gmail.com
Marvin
Hi Luke,
Am Freitag 22 Oktober 2010, um 12:42:35 schrieb Luke Kenneth Casson Leighton:
> .(..)
>
> ok i found its mailing list archives (manually) - they're empty.
> reporting a bug is not possible because of "you must have an ac100
> kernel configured" bullshit launchpad logic.
the mailing list archive is empty, because there were no posts yet ...
> the tree appears to have toshiba ac100 patches but on closer
> inspection it appears instead, at a cursory glance, to be nv-patches
> not ac100 patches.
The patches under
http://bazaar.launchpad.net/~ac100/+junk/toshiba-patches/files
are the patches against the eclair-9.12.12 branch from git://nv-
tegra.nvidia.com/linux-2.6.git which generate the same kernel source tree like
the full one submitted by Toshiba. I didn't wanted to push another tree
through my limited DSL line.
There is a full git cloneable tree at git://gitorious.org/ac100/kernel.git +
some patches if you prefer this one.
> so... i believe that the project could benefit greatly from knowing
> of the existence of the official gpl linux kernel code released
> officially by toshiba, but it is somewhat ... challenging shall we say
> to inform them of that fact.
For some reasons, Toshiba didn't wanted to make the GPL sources downloadable
from their site. They instead prefered to send CD-ROMs with the source to
everyone who asked for it.
It's great to see more and more dev's interested in this machine.
Unfortunately, the kernel source from Toshiba is not very suitable for
development for serveral reasons. This includes that the patches cannot be
forward ported to a current kernel. Also they would still depend on
proprietary user space programs (which may also not run on newer kernels), so
personally I see this patches only as some kind of reference.
IMHO, the tegra tree at android.git.kernel.org is more interesting, as it
seems to be written from scratch with "mainline mergeability" in mind, but
there is of course no support for the ac100 yet.
On the other hand, there is support for the "harmony" reference board included
(maybe limited, I don't know the current status) and the AC100 seems to be
very similar to "harmony". So this tree is probably a better place (but more
stony to start with). The tree at
http://bazaar.launchpad.net/~ac100/+junk/linux-tegra-2.6.36/files is based upon
this branch and has some board files (for paz00, ac100 board name) added, but
is does not boot.
Personally, I don't have the knowledge to do the port myself, but everyone is
welcome to join the development, maybe you want to add yourself to the ac100
team on launchpad and start hacking.
Greetings,
Marvin
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Maciej Grela
> On Tue, Oct 19, 2010 at 11:37 PM, Luke Kenneth Casson Leighton
> <lk...@lkcl.net> wrote:
>
> i haven't received a "thank you" or any kind of acknowledgement from
> you, so i do not know if you have understood the above repeated
> information. please firstly let me apologise for assuming that you're
> sufficiently unintelligent to understand why i repeated that
> information, but my concern is genuine in that i'd like you to be able
> to gain access to the source code. now let me spell it out for you.
Thank you both for the source and for pointing out my obvious mistake.
I didn't have much time to deal with this in the last days, that is
why I'm answering now.
--
Maciej Grela
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTikEeSde-i1AC1WhE...@mail.gmail.com
Luke Kenneth Casson Leighton
> Thank you both for the source and for pointing out my obvious mistake.
> I didn't have much time to deal with this in the last days, that is
> why I'm answering now.
ahh, no problem. i just like people to know that their question's
been answered. good luck with the source. l.
--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/AANLkTim_pFDsq4PcCAKbN...@mail.gmail.com