[ofa-general] Re: New proposal for memory management

[Ralph Campbell]

Correct. This is what I was referring to when I said
hardware support would be needed to make use of the 2.6.27
page mapping callback notifications. I don't know of any
current HCAs which support this sort of dynamic mapping.

On Wed, 2009-04-29 at 13:16 -0700, Ted H. Kim wrote:
> Unless you went the whole way of not really pinning
> and allowing paging of registered memory.
> Of course before anything got paged out, you would have
> to call the HCA driver to mark said pages "not present".
> And if you hit such pages when processing a work
> request in the HCA, you would have to take a page fault,
> get the page back and resume the work request
> (and probably hold off the remote side with an RNR_NAK
> in the meantime).
> 
> This was a lively topic of conversation a long time ago
> in IBTA ... and probably a bigger patch to the world
> than people were contemplating :-).
> 
> -ted
> 
> 
> > This would be nice except that it is impossible.
> > The whole idea behind virtual memory is that the application uses
> > virtual addresses which are only temporarily mapped to physical
> > memory by the OS. If the OS allowed a user to lock the
> > virtual to physical mapping for all of physical memory, the OS
> > wouldn't be able to run any other processes, do I/O, etc.
> > The process can have a larger virtual address space than the
> > available physical memory so no fixed mapping is possible.
> > To prevent this, the application has limits imposed by the OS
> > (type "ulimit -l").
>