Nishanth' tech rambles

[2013-03-06 update]: after a few years of letting things rot, I just managed to pick up my old binaries and gave it a dry run: https://github.com/nmenon/beagle-recover  has the original working binaries since the google groups link no longer work. NOTE: I HAVE NOT TRIED WITH ANY OF THE LATEST SOURCES. I do not have plans to do so in the future as well  - an quick try indicated things are broke :( [Original post] http://groups.google.com/group/beagleboard/web/beagle_recover.tar.bz2 has the recovery package ready to go for ubuntu 8.04 32bit OS: download this, run sudo ./recover_beagle.sh - as shown in the video... Steps summary: 1. first shutdown the board. 2. run 'sudo recover_beagle.sh /dev/ttyUSB0' in a terminal -> you need super user permissions(sudo permissions) to run this script. 3. Keep the pushbutton S1pressed while powering on the board -> there are two white push button switches, use the one closer to the edge of the board NOTE: if your nand is com...

After using a bunch of time getting firefox to upload from my ubuntu 8.10 box to here , I got an util which told me to use my code.google.com password instead of my google passwd.. Grrrr... anyway, couple of my sleep hours later, u-boot utils rev 0.2 "Jumpin' Monkey" is uploaded (the code name was made up as i typed ;))

for some time now, I have gone with the belief that the nand ecc imlementation such as this which I helped write good enough for nand devices, but Vimal Singh pointed me to these two articles: http://lists.infradead.org/pipermail/linux-mtd/2008-August/022611.html http://lists.infradead.org/pipermail/linux-mtd/2008-August/022613.html Essentially: If we do just old_ecc^new_ecc to detect/correct ecc errors - the logic handles the case of bit flip error at 63rd bit(as e.g.) faultily. I was completely unaware of this weakness of the algo.. thanks Vimal for pointing me here.. I have plans of setting up a bit of code to verify this (essentially hack up the current drivers by overriding the read path ;) ).. will post if I can proove(I suspect this would be the case) this is the case with gpmc..

Tony Lindgren, Linux OMAP maintainer announced that source.mvista.com linux omap copy would not be used anymore.. It has been more than three years since the discussion on hosting the kernel tree at mvista site .. Thank you mvista for hosting all our pulls to date.. from now on look here (actually it was there since 2006, though I will miss doing " linux omap git " on google).

Kinda got a few questions on how omap uboot utils works.. so here is my attempt to try and summarize what omap3 TRM says(for details TRM is the best place to go - yes, it is beautifully written ;) ): ARM is a useless unless we have some software to run on it, so OMAP has a special piece of ROM code which gets the initial control of things. Now, ROM code is capable of: a) booting off a memory device such as NOR, NAND, OneNAND, MMC etc.. kind of devices. b) booting off a peripheral such as UART, USB etc.. Before anyone asks, yep ROMCode is a software code within OMAP. And, yep, and yeah it's got all kind of drivers for devices it supports, and further yeah, it is a tiny optimized code and finally NO - you cant modify it since it is R ead O nly M emory. Now, in the world of OMAP s/w guys, (a) is called memory boot and (b) is called peripheral boot.. Now, how does ROM Code know what kind of device you have and which mode you want to boot from? This is called the SYSBOOT(again read th...