You have to write it on a usb-stick using "dd if=rootfs.ext2 of=/dev/sdb".
Also, you are using the wrong image, the only images with X are my builds and I use only tar.bz2.

The ADC works perfectly with ndless, but not my linux kernel driver..
What am I doing wrong?
My changes in nspire_mmio.h:
#define NSPIRE_ADC_PHYS_BASE 0xC4000000
#define NSPIRE_ADC_SIZE 0x000001DF
#define NSPIRE_ADC_VIRT_BASE 0xfed7ee21 //0x1df before BOOT1
common.c:
struct map_desc nspire_io_regs[] __initdata = {
IOTABLE_ENTRY(ADC),
IOTABLE_ENTRY(APB),
IOTABLE_ENTRY(BOOT1),
IOTABLE_ENTRY(INTERRUPT),
};
Then im accessing it with:
adc[0] = readl(IOMEM(NSPIRE_ADC_VIRT_BASE + 0x110));What am I doing wrong?

X is a bit too large to use as initrd.
http://dl.dropbox.com/u/105478372/zImage.tns
http://dl.dropbox.com/u/105478372/rootfs.tar.bz2
I posted these URLs somewhere in this thread, but I'm too lazy to search.


Edit: With ioremap it works.. IDK why, but I'll make a patch.
Edit2: I can't use floats inside the kernel: "Undefined reference to '__aeabi_f2d'" -> Wrong toolchain?
Edit3: Ok, moved calculation of voltages to the userspace :-/

And what is in the second RAM?

There's the SRAM (flip-flops, fast) and the SDRAM (slow, has to be refreshed dynamically)

But how knows nClock the right time after any softreboot (linucx, chrash on calc etc.)?

I think it CAN be that linucx doesn't override that value in RAM, but that's random.
The ram doesn't get cleared by linux.

BTW: We should choose a new name, linucx is outdated and wrong. Maybe linspire?

BTW: We should choose a new name, linucx is outdated and wrong. Maybe linspire?

Nope, that's the name of a commercial Linux distro

SRAM is the normal operating RAM like in a PC,

Basically all RAMs of "large" capacity (beyond several hundreds of KBs, several MBs) are SDRAMs, because they're denser (significantly fewer transistors per bit)

I think, but what's in the other one?

On the Clickpads and Touchpads, the screen buffer, for instance; but TI's OS stores some other stuff there, because the calculator usually flat out crashes, after exiting a Ndless program, if you draw past the end of the screen area.

You have to write it on a usb-stick using "dd if=rootfs.ext2 of=/dev/sdb".
Also, you are using the wrong image, the only images with X are my builds and I use only tar.bz2.

The ADC works perfectly with ndless, but not my linux kernel driver..
What am I doing wrong?
My changes in nspire_mmio.h:

Code: [Select]

#define NSPIRE_ADC_PHYS_BASE 0xC4000000
#define NSPIRE_ADC_SIZE 0x000001DF
#define NSPIRE_ADC_VIRT_BASE 0xfed7ee21 //0x1df before BOOT1
common.c:

Code: [Select]

struct map_desc nspire_io_regs[] __initdata = {
IOTABLE_ENTRY(ADC),
IOTABLE_ENTRY(APB),
IOTABLE_ENTRY(BOOT1),
IOTABLE_ENTRY(INTERRUPT),
};
Then im accessing it with:

Code: [Select]

adc[0] = readl(IOMEM(NSPIRE_ADC_VIRT_BASE + 0x110));What am I doing wrong?

Just make the size SZ_4K. You're also doing unaligned reads because addresses start at 0xfed7ee21 which isn't 4 byte aligned.

Is SZ_4K =1 because the size has to be in pages?
I'll change it, test it and send you a new patch.
Does X and directfb work on click and touchpads?
If not, you have to add cx_clcd_check to your classic clcd panel struct.

No, it means 4096 bytes. It's a lot easier than calculating everything exactly right.

Edit: X doesn't appear to work, I'll try including the clcd_check and try again

it didn't work because the MMU can only map whole pages, not bytes,
like it is on x86 with PAGing?

That's possible.