[Question] detach_vmas_to_be_unmapped

[akuster <akuster@mvista.com>]

Hello,

I am a bit confused on why the vm start and end address where swapped in 
arch_unmap_area and arch_unmap_area_topdown functions after the official 
Avoiding mmap fragmentation ( 1363c3cd8603a913a27e2995dccbd70d5312d8e6 ) 
patch.

Prior to this patch arch_unmap_area() used area->vm_start and 
arch_unmap_area_topdown used area->vm_end after the patch the following 
change showed up.

if (mm->unmap_area == arch_unmap_area)
     addr = prev ? prev->vm_start : mm->mmap_base;
else
     addr = vma ?  vma->vm_end : mm->mmap_base;

Using Wolfgang Wander's leakme.c test, I get the same results seen with 
his original "Avoiding mmap fragmentation" patch as I do after swapping 
  the start & end address in the above code segment.

Is it just a typo in the official patch or am I missing the much bigger 
picture?

Note: I have only run this on an Pentium4 with 1GiG memory.

TIA

Armin