Geekmaster's Playground

From MobileRead
Jump to: navigation, search

Contents

[edit] Introduction

This is a little area where I will document my Kindle research, experiments, and other related fun stuff.


geekmaster

I have other websites at my disposal, but this one comes with a built-in editor.

Some of the ideas presented here will be works-in-progress, and as such will change and evolve, and the as the content evolves it will get adjusted, adapted, refactored, rearranged. Some content may move to separate pages, or even be deleted if I decided so.

If you stumbled upon this page, and you wish to add a comment, please feel free to do so. If you see a spelling or grammar error, you may correct it. If you disagree with an IDEA, please do not change my content, but add a short description of why you think it is in error, with some supporting evidence (such as a link to related content).


Thank you for your cooperation.


[edit] Build Environments

My first interest is getting some native mode compiler tools up and running that work on all kindles, are small and simple, and that do not require loop mounting a file system image.

Tinycc (also called tcc) seems to fit my wishes quite nicely: http://www.tinycc.org http://repo.or.cz/w/tinycc.git

There was a version supplied by Luigi Rizzo, which he said was missing headers, but it appears to be missing libraries as well. I would prefer to have a static-linked version that has no external dependencies except the libtcc library (also static-linked) to allow embedded compilation inside applications.


[edit] Animation

Although it may seem strange at first, I want to have some degree of interactive graphics and animation on my kindles, including the ability to display live video. I have had some degree of success in my experiments into this. Of course, the newer eInk pearl display does a much better job.

I have discovered that the kindles can update the display MUCH faster with it contains only black and white pixels. Gray pixels require extra time to change. I have tried various forms of dithering to simulate grayscale, and I have decided that a Bayer ordered-dither causes the least time-domain artifacts due to causing minimal pixel content changes for moving objects.


[edit] Misc. Notes

This is from touch diags /etc/upstart/diags:

 /usr/local/sbin/dropbear   <== this is interesting. diags sshd?

^^^ 192.168.15.200:/scratchbox/nvf/factory NFS mounted at /mnt/gen3/factory? ^^^^ when diags usbnet on, nmap says that touch listens on port 6 ??? ...

 # Not set the cpu governor speed to ondemand for V&I station
 echo ondemand > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor

...

 # start performance measurements
 # 30ms sample rate (1% CPU) buffered until we send SIGTERM in framework.conf
 # TODO - make sure production loses /usr/local/bin to disable this
 [ -x /usr/local/bin/perf ] && perf -s30 -b/tmp/perf.out &

...

see: Geekmaster's Playground - Select Boot

Personal tools
Namespaces

Variants
Actions
Navigation
MobileRead Networks
Toolbox