On Mar 3, 2005, at 5:00 AM, Flemming Munk Jensen wrote:

>> which would mean that the 1.2b6 WRAP image with its 7995392
>> bytes should fit. I intentionally made the CF root filesystem
>> slightly less than 8 million bytes in the hope that it'd still fit on
>> an 8 MB CF card, but apparently that didn't work out (maybe due to
>> CHS?). Can you determine how many sectors your 8 MB CF card actually
>> has?
>>
> This is what the bios on my VIA EPIA motherboard tells me about the CF 
> Card:
>
> Capacity : 8 MB
> Cylinder : 245
> Head : 2
> Precomp : 0
> Landingzone : 244
> Sectors : 32

Your card is (slightly larger than) 8MB

245 * 2 * 32 =15680
15680 * 512 = 8028160

However, it looks like your CF card reserves the last full cylinder as 
a "landing zone".

244 * 2 * 32 = 15616
15616 * 512 = 7995392

Which is, incidentally the exact size of the 1.2b6 image, so if 
everything is working (physdiskwrite, the CF mechanism on the windows 
box
used to run physdiskwrite, the CF card, etc), it should fit, and be OK.

So now I wonder if physdiskwrite has a little fencepost error in it (is 
the source available?) or perhaps
one of the last sectors on your CF is bad (and you've never hit it 
before)?

If you've got access to a linux/bsd/solaris/OS X box try this:

zcat mono-1.2b6.img | dd of=/dev/XXX conv=sync

where '/dev/XXX' is the (raw/character) device of the CF reader 
attached.  You'll have to know something about *nix
to find this information.    On my USB-connected widget, its /dev/sdc.

jim