On Tue, Jan 05, 2016 at 02:59:03PM +0100, Michal Hocko wrote: > > 3) vmalloc() is for large allocations. They will be page-aligned, > > but *not* physically contiguous. OTOH, large physically contiguous > > allocations are generally a bad idea. Unlike other allocators, there's > > no variant that could be used in interrupt; freeing is possible there, > > but allocation is not. Note that non-blocking variant *does* exist - > > __vmalloc(size, GFP_ATOMIC, PAGE_KERNEL) can be used in atomic > > contexts; it's the interrupt ones that are no-go. The last sentence I'd put into that part was complete crap... > It is also hardcoded GFP_KERNEL context so a usage from NOFS context > needs a special treatment. ... in part because of this. GFP_ATOMIC __vmalloc() will be anything but, and the only caller passing that is almost certainly bogus. As for NOFS/NOIO, I wonder if we should apply that special treatment inside __vmalloc_area_node rather than in callers; see the current thread on linux-mm for details... Another interesting issue is __GFP_HIGHMEM meaning for kmalloc and __vmalloc resp. (should never be passed to kmalloc, should almost always be passed to __vmalloc - the former needs pages mapped in kernel space, the latter probably never needs a separate kernel alias for the data pages, to such degree that I'm not sure if we shouldn't _force_ __GFP_HIGHMEM for data pages allocation in __vmalloc_area_node()) > > 4) if it's very early in bootstrap, alloc_bootmem() and friends > > may be the only option. Rule of the thumb: if it's already printed > > Memory: ...../..... available..... > > you shouldn't be using that one. Allocations are physically contiguous > > and at that point large physically contiguous allocations are still OK. Probably needs at least some discussion of memblock vs. bootmem APIs. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>