On Mon, May 25, 2020 at 05:19:11PM +0300, Konstantin Khlebnikov wrote: > Tool 'page-types' could list pages mapped by process or file cache pages, > but it shows only limited amount of state exported via procfs. > > Let's employ existing helper dump_page() to reach remaining information: > writing pfn into /sys/kernel/debug/dump_page dumps state into kernel log. > > # echo 0x37c43c > /sys/kernel/debug/dump_page > # dmesg | tail -6 > page:ffffcb0b0df10f00 refcount:1 mapcount:0 mapping:000000007755d3d9 index:0x30 > 0xffffffffae4239e0 name:"libGeoIP.so.1.6.9" > flags: 0x200000000020014(uptodate|lru|mappedtodisk) > raw: 0200000000020014 ffffcb0b187fd288 ffffcb0b189e6248 ffff9528a04afe10 > raw: 0000000000000030 0000000000000000 00000001ffffffff 0000000000000000 > page dumped because: debugfs request > > With CONFIG_PAGE_OWNER=y shows also stacks for last page alloc and free: > > page:ffffea0018fff480 refcount:1 mapcount:1 mapping:0000000000000000 index:0x7f9f28f62 > anon flags: 0x100000000080034(uptodate|lru|active|swapbacked) > raw: 0100000000080034 ffffea00184140c8 ffffea0018517d88 ffff8886076ba161 > raw: 00000007f9f28f62 0000000000000000 0000000100000000 ffff888bfc79f000 > page dumped because: debugfs request > page->mem_cgroup:ffff888bfc79f000 > page_owner tracks the page as allocated > page last allocated via order 0, migratetype Movable, gfp_mask 0x100dca(GFP_HIGHUSER_MOVABLE|__GFP_ZERO) > prep_new_page+0x139/0x1a0 > get_page_from_freelist+0xde9/0x14e0 > __alloc_pages_nodemask+0x18b/0x360 > alloc_pages_vma+0x7c/0x270 > __handle_mm_fault+0xd40/0x12b0 > handle_mm_fault+0xe7/0x1e0 > do_page_fault+0x2d5/0x610 > page_fault+0x2f/0x40 > page last free stack trace: > free_pcp_prepare+0x11e/0x1c0 > free_unref_page_list+0x71/0x180 > release_pages+0x31e/0x480 > tlb_flush_mmu+0x44/0x150 > tlb_finish_mmu+0x3d/0x70 > exit_mmap+0xdd/0x1a0 > mmput+0x70/0x140 > do_exit+0x33f/0xc40 > do_group_exit+0x3a/0xa0 > __x64_sys_exit_group+0x14/0x20 > do_syscall_64+0x48/0x130 > entry_SYSCALL_64_after_hwframe+0x44/0xa9 > > Signed-off-by: Konstantin Khlebnikov <khlebnikov@xxxxxxxxxxxxxx> Looks useful to me: Acked-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx> -- Kirill A. Shutemov