This post might not apply to those whos kernal versions already work correctly with their HDDs.

I have had a problem booting my menuet kernal and ramdisk from my HDD on my toshiba laptop.. ever since version 75 (matts gridworks release)
After a few decent hours of deliberation with a file comparison utility and some kernal code swaping... (between 76 beta and 78pre1). I discovered that one of the main problems was simply the filename... kernal 78 tries to load msetup.exe as the image... i know that this is the installer... this is trivial... for those of us with meosload.com we can always rename the file...
But now is a good time to decide bewteen a true image file and a hacked executable...

the filename was trivial... so it wasnt the main problem...

It still wouldnt boot with the name change, later i found that the fat32 code that loads the ramdisk image from the HDD didnt work properly. I exchanged the older functions from version 76 into version 78 and it booted properly off the HDD... however... MFAR couldnt read the HDD properly... this might have to do with code conflicts...

So... i cannot say EXACTLY what causes the problem... but it has to do with the code that loads the ramdisk properly...

Note that all versions boot from my floppy disk perfectly fine.. since this uses seperate code to load the ramdisk.

I COULD botch up some fix to this... but id rather not muss things up with some kind of second version/release... I just thought id give a heads up to anyone doing bug fixing in the kernal... if its needed i can point out specific procedures and line numbers of what i suspect to be problem code.
Many thanks,
Wishing

Hi Wishing,

We have tested and looked @ lots of code lately.

We try to keep the functionality high with the MOST PCs we can.

( the more that see our demo's / distros = more interest = more code.)

Optimizing/Standardizing the kernel / FAT 32 code seems to be a work in progress. From PC to PC Ivan's FAT32 works well. New.78p2 is working on more machines than earlier releases.


Any Info you've gathered or your results might help in a solution.

BTW We like the tinypad mods alot!

When you get a chance send your info to us.

Thanks, Mark & Matt

There must be some specific error that occurs before the fat32 chain is built.

I never get to see any text written.. so that means i never get to the first boot_log. The ONLY loop before that boot log (since no multitasking is occuring this must be true) is within the HDD ramdisk loading...
lines 404: jb hdbootl1

that just jumps back up again... the only reason it would do so is if it never got the correct amount of bytes loaded...

lines 401 through 403 are the ones i suspect to be in error
(perhaps this is because Menuet.img is a full disk image and msetup.exe is a different size?)

In the current kernal i have running (76 beta) it uses rdload.inc instead of the code inside the kernal (i can see how it was simply moved over though)

compare lines 38-40 of rdload.inc (76 beta) to lines 401-404 in the kernal of 78pre1.

In these lines the comparison is made... in the old version it said
cmp ecx,1474560/512+1
this makes sense...
however in the new version it has
cmp esi,1474560/512+41-1
Its reading extra bytes...
I think these bytes are the .exe top part of msetup.exe...
so... with this in mind...
I will try to use ONLY msetup.exe on my laptop along with the new kernal...
if this does not work... then something else is wrong...

regardless of this fact... as ive said before... a standerdized ramdisk image file should be agreeed upon... should it have installer code jammed atop it?
or should it be a true image file? Im inclined to use the latter... since true images can be burned to disks etc.

Anyways... ill try what i said.. if anyone else has any ideas/fixes tell me...

Im glad people like my tiny pad improvements.. i have plenty more instore
Wishing

There must be some specific error that occurs before the fat32 chain is built.

I never get to see any text written.. so that means i never get to the first boot_log. The ONLY loop before that boot log (since no multitasking is occuring this must be true) is within the HDD ramdisk loading...
lines 404: jb hdbootl1

that just jumps back up again... the only reason it would do so is if it never got the correct amount of bytes loaded...

lines 401 through 403 are the ones i suspect to be in error
(perhaps this is because Menuet.img is a full disk image and msetup.exe is a different size?)

In the current kernal i have running (76 beta) it uses rdload.inc instead of the code inside the kernal (i can see how it was simply moved over though)

compare lines 38-40 of rdload.inc (76 beta) to lines 401-404 in the kernal of 78pre1.

In these lines the comparison is made... in the old version it said
cmp ecx,1474560/512+1
this makes sense...
however in the new version it has
cmp esi,1474560/512+41-1
Its reading extra bytes...
I think these bytes are the .exe top part of msetup.exe...
so... with this in mind...
I will try to use ONLY msetup.exe on my laptop along with the new kernal...
if this does not work... then something else is wrong...

regardless of this fact... as ive said before... a standerdized ramdisk image file should be agreeed upon... should it have installer code jammed atop it?
or should it be a true image file? Im inclined to use the latter... since true images can be burned to disks etc.

Anyways... ill try what i said.. if anyone else has any ideas/fixes tell me...

Im glad people like my tiny pad improvements.. i have plenty more instore
Wishing

I've the same problem with my ACER laptop