On Mon 04-02-13 23:48:46, Ric Mason wrote: > Hi Andi, > On Thu, 2013-01-31 at 11:52 +0100, Andi Kleen wrote: > > On Thu, Jan 31, 2013 at 02:23:10AM -0600, Ric Mason wrote: > > > Hi all, > > > > > > It seems that Andi's "Support more pagesizes for > > > MAP_HUGETLB/SHM_HUGETLB" patch has already merged. According to the > > > patch, x86 will support 2MB and 1GB huge pages. But I just see > > > hugepages-2048kB under /sys/kernel/mm/hugepages/ on my x86_32 PAE desktop. > > > Where is 1GB huge pages? > > > > 1GB pages are only supported under 64bit kernels, and also > > only if you allocate them explicitely with boot options. > > I am curious about how can buddy system alloc 1GB huge pages? the most > order buddy system supports is 10. Could you explain to me? Have a look at setup_hugepagesz & hugetlb_nrpages_setup (in x86 arch code). A short answer is. Bootmem is used for allocation > MAX_ORDER. -- Michal Hocko SUSE Labs -- 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>