Messages in this thread |  | | Date | Fri, 15 Sep 2000 09:54:48 -0500 | From | Eli Carter <> | Subject | Re: Fixed addresses for various architectures |
| |
Russell King wrote: > Eli Carter writes: > > I have a Gnome Dia document in which I've tried to lay out the fixed > > memory locations for arm-linux and the mapping of > > ioremap<->virt<->phys<->bus. It's not perfect, but if you want a copy, > > email me. > > Note that all of this is dependent on many things: machine type, binary > type etc.
I do realize it is machine type dependant... I am not familiar with the binary type dependancies--can someone point me to information/explaination about that?
The virtual memory mapping looks like it is (or maybe can be?) generalized in terms of kernel constants (#defines). My knowledge is based on your arm-linux, let me know if things are different for other archs or if I just got something plain wrong...
0 to TASK_SIZE is user process memory space (with the 0 page accesses trapped or something...) PAGE_OFFSET is the start of physical memory, with memory_start to memory_end being what's left after the kernel takes what it wants. VMALLOC_START to VMALLOC_END is the memory space reserved for ioremapped addresses (On the ARM, that still leaves a blank above e0000000... what's up there?)
These #defines seem to appear in most/all architectures, just with different values. What of the above is bogus? Where can I get a more accurate and complete understanding of this?
C-ya,
Eli --------------------. "To the systems programmer, users and applications Eli Carter | serve only to provide a test load." eli.carter@inet.com `---------------------------------- (random fortune) - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
|  |