Toshiba AC100

[pibach2]

macau, any news on your Arch installation? Does it work with suspend and speaker sound? Work on Ubuntu Oneiric seem to mature meanwhile, I got 2.6.38-1000-ac100 kernel through simple update. Still no suspend (but blnking LED) and sound is only via headset.

[macau]

with last commits(a96b4f5b8713dcc07192a2f26d872beeb4ba8c36) suspend work fine. sound still in headphones only. touchpad and kb dont lost now.

[pibach2]

with last commits(a96b4f5b8713dcc07192a2f26d872beeb4ba8c36) suspend work fine. sound still in headphones only. touchpad and kb dont lost now.

hm, I cannot identify that commit, is it in here?
http://gitorious.org/~marvin24/ac100/ma ... 100-2.6.38

My Ubuntu install still gets performance bugs, i.e., when high load on Ram it becomes unresponsive. I suspect mmcqd and some deadlock problem. I am on the internal MMC and ext4, don't know if that matters. Maybe nilfs2 is better? What about that in your Arch install?

[macau]

u must pull marvin's tree to last commits. ubuntu use very old kernel i use arch on external usb drive and internal emmc with ext4, and can say that performance more sensetive than in ubuntu

[benob]

I successfully installed archlinux/arm on my ac100. Wifi/suspend work fine, but I didn't try sound.

Any idea or pointers for how to change brightness?
How about detecting lid close to suspend to ram?
Any fix for the evbug reports and weird keyboard behavior (arrow keys self-repeating)?
Any kernel around with zram support?

[Shu]

Any idea or pointers for how to change brightness?
How about detecting lid close to suspend to ram?
Any fix for the evbug reports and weird keyboard behavior (arrow keys self-repeating)?
Any kernel around with zram support?

1.

Code: Select all

#!/bin/bash

CUR=`cat /sys/class/backlight/pwm-backlight/brightness`
   
case "$1" in
plus)
    if [ "$CUR"  = 254 ]; then
        exit 0
    fi
    NEW=`calc "$CUR"+10`
    ;;
minus)
    if [ "$CUR"  = 30 ]; then
        exit 0
    fi
    NEW=`calc "$CUR"-10`
    ;;
*)
    echo "usage: $0 {plus|minus}"     
esac
/bin/echo -n $NEW > /sys/class/backlight/pwm-backlight/brightness


2. Not ready yet
3. rmmod evbug
or add "blacklist evbug" to /etc/modprobe.d/modprobe.conf
4. Zram support was recently added to the marvin24s-kernel

[pibach2]

I successfully installed archlinux/arm on my ac100. Wifi/suspend work fine,

Is it true, that resume works only with Android 2.1 bootloader while 2.2 doesn't?
So which bootloader are you using?

Also could you please do some test regarding memory lock problem? Start for example Chromium with 20 webpages and LibreOffice, then do some package update & upgrade in the background. Do you see mmcqd consuming the CPU? And freezing GUI?

[benob]

Thank you very much, Shu.

Pibach2, I did install it on top of Android 2.2, using partition 6 (5 was with Android 2.1).
As for memory lockups, I didn't setup a swap, so chromium got killed after 10 or so tabs open and everything else continued smoothly.

I have another question. I use boot-marvin24s-2.6.38-095bb47 and it only exposes /dev/mmcblk0 (no partitions) for internal storage. Can I create a partition table on it or is it going to kill the kernel partition?

tegrapart from cmdline: recovery:300:a00:800,boot:d00:1000:800,mbr:1d00:200:800

Also, anyone was successful using abootimg to flash a kernel from within the ac100?

[pibach2]

Pibach2, I did install it on top of Android 2.2, using partition 6 (5 was with Android 2.1).

In the Ubuntu Wiki now there is a new hint:

"Kernel
As of early September 2011 a 2.3.38-chromeos kernel fork is used in Ubuntu and included in the Oneiric archives. It supports suspend/resume on machines that did not upgrade to the Android 2.2 system and bootloader. "

seemingly suspend and Android 2.2 bootloader have a conflict.
So I need to reflash Android 2.1 bootloader back on partition as described here: https://lists.launchpad.net/ac100/msg00311.html

As for memory lockups, I didn't setup a swap, so chromium got killed after 10 or so tabs open and everything else continued smoothly.

ok. But I think these memory lock problems are not only related to swapping. But any concurrent read+write (!) access to the eMMC. Nobody has that?

And does the new zram support help much?

[macau]

"Kernel
As of early September 2011 a 2.3.38-chromeos kernel fork is used in Ubuntu and included in the Oneiric archives. It supports suspend/resume on machines that did not upgrade to the Android 2.2 system and bootloader. "

seemingly suspend and Android 2.2 bootloader have a conflict.

i have 2.2 bootloder and working suspend, so there are ubuntu-specific bug about ram:
try opening 20 tabs at x86 machine with 512M of ram in chromium and see how much ram eat that monster