[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: APEX and Debian

On Sun, Jul 16, 2006 at 03:12:32AM +0200, Kevin Price wrote:
> Thanks a lot for nslu2-utils. I think that most users prefer not to flash the
> slug too often, just like I. I know that flash EEPROMS cannot stand infinetely
> many flashes 

Well, intel promises[1] 100K minimum erase cycles for the flash chips in
slug, so I wouldn't be overly worried. 

> and I know that if I flash something bad then my slug goes to
> NAS-heaven. 

Technically flashing something bad should not take your slug
to NAS-heaven. nslu2-utils does not touch redboot, and thus
if you flash a bad kernel+initramfs you should still have
redboot console available.

However a bug in any of the tools in the flashing chain that
would cause wrong flash blocks getting overwritten could indeed
brick it, so some carefulness is a good idea..

> Are you convinced that most users think like I do or should I write more arguments?

Adding usb and mass storage support to a bootloader would be a 
massive job. Maybe kexec[2] on arm[3] would be a better way to pursue
loading kernel from HD. Thus there would be a minimal kernel on
flash, and a initramfs dedicated to finding and kexecing 
the "real" kernel from HD.


[1] http://www.intel.com/design/flcomp/datashts/290667.htm
[2] http://www.xmission.com/~ebiederm/files/kexec/README
[3] http://www.rpsys.net/openzaurus/patches/kexec-arm-r1.patch

Reply to: