--- Begin Message ---
On Sun, Feb 06, 2005 at 01:24:26AM +0100, Sven Mueller wrote:
> Kurt Roeckx wrote on 03/02/2005 18:24:
> >
> >Actually, you can use 4GB of address space for userspace.
>
> For 32bit userspace or also for 64bit userspace?
>
> Or in other words: With a 64bit (amd64) kernel, where are the actual
> limits for:
> 32bit Apps
4GB, but you can change it to 3GB if you want.
> 64bit Apps
/proc/cpuinfo says:
address sizes : 40 bits physical, 48 bits virtual
And from /usr/include/asm/processor:
/*
* User space process size: 512GB - 1GB (default).
*/
#define TASK_SIZE (0x0000007fc0000000)
Which is only 39 bit.
There is also this in page.h:
/* See Documentation/x86_64/mm.txt for a description of the
* memory map. */
#define __START_KERNEL 0xffffffff80100000
#define __START_KERNEL_map 0xffffffff80000000
#define __PAGE_OFFSET 0x0000010000000000 /* 1 << 40 */
#define __PHYSICAL_MASK_SHIFT 40
#define __PHYSICAL_MASK ((1UL << __PHYSICAL_MASK_SHIFT) - 1)
#define __VIRTUAL_MASK_SHIFT 48
#define __VIRTUAL_MASK ((1UL << __VIRTUAL_MASK_SHIFT) - 1)
And from Documentation/x86_64/mm.txt:
The paging design used on the x86-64 linux kernel port in 2.4.x
provides:
o per process virtual address space limit of 512 Gigabytes
o top of userspace stack located at address 0x0000007fffffffff
o PAGE_OFFSET = 0xffff800000000000
o start of the kernel = 0xffffffff800000000
o global RAM per system 2^64-PAGE_OFFSET-sizeof(kernel) = 128 Terabytes - 2 Gigabytes
Note that that doc is about the 2.4 kernel and they're
already talking about removing the 512GB per process limit. I
think it shouldn't be a problem moving it to the 47 bit limit
(512TB), but that's probably going to take some work.
Kurt
--- End Message ---