Re: [PATCH 2.6.21] cramfs: add cramfs Linear XIP

[Carsten Otte <cotte@de.ibm.com>]

Nick Piggin wrote:
> Carsten Otte wrote:
>> The current xip stack relies on having struct page behind the memory 
>> segment. This causes few impact on memory management, but occupies 
>> some more memory. The cramfs patch chose to modify copy on write in 
>> order to deal with vmas that don't have struct page behind.
>> So far, Hugh and Linus have shown strong opposition against copy on 
>> write with no struct page behind. If this implementation is acceptable 
>> to the them, it seems preferable to me over wasting memory. The xip 
>> stack should be modified to use this vma flag in that case.
> 
> I would rather not :P
> 
> We can copy on write without a struct page behind the source today, no?
> What is insufficient for the XIP code with the current COW?

I've looked at the -mm version of mm/memory.c today, with intend to 
try out VM_PFNMAP for our xip mappings and replace nopage() with fault().
The thing is, I believe it does'nt work for us:
  * The way we recognize those mappings is through the rules set up
  * by "remap_pfn_range()": the vma will have the VM_PFNMAP bit set,
  * and the vm_pgoff will point to the first PFN mapped: thus every
  * page that is a raw mapping will always honor the rule
  *
  *      pfn_of_page == vma->vm_pgoff + ((addr - vma->vm_start) >> 
PAGE_SHIFT)

This is, as far as I can tell, not true for our xip mappings. Ext2 may 
spread the physical pages behind a given file all over its media. That 
means, that the pfns of the pages that form a vma may be more or less 
random rather than contiguous. The common memory management code 
cannot tell whether or not a given page has been COW'ed.
Did I miss something?

so long,
Carsten