Re: [RFCv2][PATCH 1/3] create slow_virt_to_phys()

2012-12-09 Thread Gleb Natapov
Copying more people. Is this approach good? The alternative would be to allocate NR_CPUS sized arrays in KVM. On Fri, Dec 07, 2012 at 04:30:23PM -0500, Dave Hansen wrote: > > This is necessary because __pa() does not work on some kinds of > memory, like vmalloc() or the alloc_remap() areas on

Re: [RFCv2][PATCH 1/3] create slow_virt_to_phys()

2012-12-09 Thread Gleb Natapov
Copying more people. Is this approach good? The alternative would be to allocate NR_CPUS sized arrays in KVM. On Fri, Dec 07, 2012 at 04:30:23PM -0500, Dave Hansen wrote: This is necessary because __pa() does not work on some kinds of memory, like vmalloc() or the alloc_remap() areas on

[RFCv2][PATCH 1/3] create slow_virt_to_phys()

2012-12-07 Thread Dave Hansen
This is necessary because __pa() does not work on some kinds of memory, like vmalloc() or the alloc_remap() areas on 32-bit NUMA systems. We have some functions to do conversions _like_ this in the vmalloc() code (like vmalloc_to_page()), but they do not work on sizes other than 4k pages. We

[RFCv2][PATCH 1/3] create slow_virt_to_phys()

2012-12-07 Thread Dave Hansen
This is necessary because __pa() does not work on some kinds of memory, like vmalloc() or the alloc_remap() areas on 32-bit NUMA systems. We have some functions to do conversions _like_ this in the vmalloc() code (like vmalloc_to_page()), but they do not work on sizes other than 4k pages. We