Mobilepro 900c and linux is possible?

OK, sorry to keep people waiting bootable kernel image with keyboard support is available for download from here

http://epiccrossings.mrtus.net/xoops/htdocs/modules/docmanager/inde...

I'll be putting the kernel patches up on the site in the next few days too which should help with development ...

very cool.! sadly i missed you on irc going to try this soon..

in answer to your question on irc: i've not done the bootable image yet, didn't have the time, and it does take more time than initially estimated. but almost there now (hopefully i'm right and the ram usage is as configurable in the image as it appears)

Hi Guys, there are any news ?

I wouldn't mind doing some testing on the Linux system ( have spare 900C's ) but I can't even get Haret running on the cmonex v1.0 Rom - should I reflash to basic 900C?

I wouldn't mind doing some testing on the Linux system ( have spare 900C's ) but I can't even get Haret running on the cmonex v1.0 Rom - should I reflash to basic 900C, or, which version of Haret should I use?
And - Haret starting Gapi - any reason?

iirc, the gapi should be disabled. haret does use gapi but somehow it doesn't work right with the wincesoft gapi.

yep. Since I could not find how to disable the GAPI, I reflashed with the standard rom.

Now I can boot Linux with the latest HARET. I'll be checking for the DiskOnChip etc. - some experience in that.

cmonex, if we'll have a linux bootloader on your Rom OS image, this should be addressed.

Anyway: It Woiks!!

To get back to CE: just enter reboot, and go through the hard reset extra's

ok, i will check for you how to disable the gapi but probably you just copy an invalid gx.dll into \windows

what kind of experience do you have with DOC? i'd be very interested to hear about that.

Flashing DOC's to change dumb Thin Clients/WBT's to Linux machines. And addressing them from Linux
see http://damnsmalllinux.org/cgi-bin/forums/ikonboard.cgi?;act=ST;f=26...

BTW- can not find the DOC yet from Linux on the 900C- has msys support been compiled in the kernel? - pretty default by now.

yes, sorry to miss you on irc cmonex - I'm not surprised to hear the CF booting takes some figuring out. Actually I just had to reflash my OS from cf, was trying some different things under linux and I somehow corrupted the flash but luckily the bootloader survived ...

So I better insert the usual warning here to testers out there ... the current linux code for MobilePro900C is very beta and could even damage your hardware ... still be brave, we'll get the bugs sorted out

Anyhow I think booting direct from CF will be a big improvement over HaRET. There's a few details I'll have to get to you about a couple of params the linux kernel expects to be passed from the bootloader - can't find the link right now but I'll get the info to you.

michelbel - good to hear the current image works for you. DOC would be great to get functional, it seems to be on the PXA255's static CS2, which means it's addressable from a base physical address 0x0800_0000, looking at the datasheet for M-systems disk on chip md3831-d32-v3-x you can work out how to address the registers etc. And in the kernel sources M-systems DOC drivers have been written as you point out - so those drivers should need not much more than a little tweaking to get them working ... I did try compiling the kernel with M-system support unmodified and it borked

Also I have Compact Flash and touchscreen close to working, I can actually mount and read and write a CF card. It turns out the touch screen is coming through the BTUART along with the keyboard so I'm reworking the keyboard driver to share with the touchscreen ... once again 'real soon now' I will have kernel patches and an image available.

Then we should be able to do exciting things like boot up cool root filesystems such as jlime's ones, or pretty much any rootfs you can put together

cheers cosmo0

Well, yes, I know how to recognize the _real_ scouts: that's those with the arrows in their back. Still, it was already worth the effort.

Michel

cosmo0: shame youve not been on irc lately

uh. that is scary, how did you manage to corrupt the OS image? i'd like to hear details.
glad you flashed bootloader 4.10 in time

to everyone: UPDATE TO BOOTLOADER 4.10 BEFORE PLAYING WITH THIS BETA.

(unless i/we get some RAM OS to boot and if the buggy bootloader can handle this task at all. right now, i don't have one to test if it can boot an image, and didn't feel like flashing one with the crap bootloader yet just to test, but soon. )

congrats on the touch screen and CF though! i should've guessed touch screen is handled there

anyway..my news: restarting work on jtagging with another friend
i didn't have time to deal with the CF booting, image was half finished when i found out i screwed up something.. have to start almost from scratch.
as i said the ram xipkernel itself boots fine, as it is already ready for xip from ram.
and yes it would be great, to be able to skip booting to CE first if you want to load linux instead.

can you give me what params it wants? anyway i guess it should be no big problem, we can first execute another bootloader code, to get ready to execute the kernel itself.

as for the m-systems DOC i don't know much about it even though i'm very interested in it. i want to access the nand (probably DOC) flash in another hpc to copy the OS image, and there is no documentation. i do have some bootloader code for it, but it'll need some playign with
so i'd like to know as much as possible about this topic. help
sorry for the off


Edited by cmonex 2007-05-29 2:13 PM

update: jtag pins measured successfully! we are working on getting a unit actually jtagged. the connector is very weird so it will be an ugly solder job. (there is a trick but i'll tell about it only if it works )
what software would you recommend for this flash chip / cpu? (the flash is the intel 28F128J3C150 as a reminder)

that is GREAT cmonex

have a look at this link to a gumstix page about JTAG. They seem to be using OCD Commander to flash the bootloader even though OCD Commander (free software) claims not to do flashing. They mention a couple other softwares too and the hardware info is very interesting, although Gumstix specific ...

When I corrupted the flash I was forcing some weak kernel code of mine to overflow and crash at the time, it sure did overflow and crash but thankfully was recoverable.


LOL, arrows removed I just want to make sure I don't release some code that could wipe people's flash so I'll do some more intensive testing before CF and touchscreen code is released ... but that said it shouldn't be too long, I know where the biggest problem in the code is and I'll make sure it can't crash the kernel like that before next release, for now use at your own risk and I recomend people

DON'T USE THE TOUCHSCREEN STYLUS

while testing the current keyboard driver as the keyboard driver will just get flooded with spurious touchscreen packets and bork your kernel unpredictably


That's what I was thinking too, I'll be in touch, irc or email soon with those parameters for you ...

cheers cosmo0

hey cosmo0: ocd commander cannot do flashing unless it is a crack i don't know of. also, i'm not buying their flash software, it costs 500$ to be able to program the flash. i need FREEWARE. and, jlink looked like it is a hardware jtag programmer. tell me if i missed something.
what do you think about jtag tools 0.5.x? my problem is that my friend probably won't install linux to use this one

why were you trying to overflow stuff?
waiting for your email i wasn't on irc since last night, will get there now

Edited by cmonex 2007-05-30 9:10 AM