On Sun, 2006-09-24 at 18:36 -0700, Jeremy Fitzhardinge wrote:

Rusty Russell wrote:

quoted

	You're thinking of it in a convoluted way, by converting to offsets
from the per-cpu section, then converting it back.  How about this
explanation: the local cpu's versions are offset from where the compiler
thinks they are by __per_cpu_offset[cpu].  We set the segment base to
__per_cpu_offset[cpu], so "%gs:per_cpu__foo" gets us straight to the
local cpu version.  __per_cpu_offset[cpu] is always positive (kernel
image sits at bottom of kernel address space).
  

We're talking kernel virtual addresses, so the physical load address 
doesn't matter, of course.

So, take this kernel I have here as an explicit example:

$ nm -n vmlinux
[...]
c0431100 A __per_cpu_start
[...]
c0433800 D per_cpu__cpu_gdt_descr
c0433880 D per_cpu__cpu_tlbstate


And say that this CPU has its percpu data allocated at 0xc100000.

That can't happen, since 0xc100000 is not in the kernel address space.
0xc1000000 is though, perhaps that's what you meant?

So, in this case the %gs base will be loaded with 0xc100000-0xc0431100 = 
0x4bccef00

A negative offset, exactly, which can't happen, as I said.

Hope that clarifies?

Confused,
Rusty.
-- 
Help! Save Australia from the worst of the DMCA: http://linux.org.au/law