From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Mon, 26 Jun 2000 16:45:34 -0500 From: Timur Tabi Subject: referenced/uptodate pages in the free list? Message-Id: <20000626215518Z131165-21002+48@kanga.kvack.org> Sender: owner-linux-mm@kvack.org Return-Path: To: Linux MM mailing list List-ID: I've written a small driver which traverses the free list and prints out a list of all the pages that are free (yes, I know it's a huge list). The code goes from contig_page_data.node_zones[1].free_area[0] to contig_page_data.node_zones[1].free_area[MAX_ORDER-1]. It traverses the free_list for each free_area. If I read everything correctly, each element in free_area[x].free_list is a pointer to a physically contiguous list of 2^x elements in the mem_map array. IOW, if free_area[2].free_list.prev points to a mem_map_t structure, then it means that that mem_map_t and the 3 following it are supposed to be all free. If I have that right, then I'm seeing something odd. My program prints out the contents of each mem_map_t pointed to be each element of free_list of each free_area of orders 0, 1, 2, and 3. I'm seeing something weird: Jun 26 16:30:57 two kernel: O2: free_area c025c8a0, prev=c127bb28, next=c129ad88, *map=0 Jun 26 16:30:57 two kernel: mem_map_t at c127bb28: phys: 8d44000 flags: 0 zone: c025c86c, phys: 8d45000 flags: 0 zone: c025c86c, phys: 8d46000 flags: 0 zone: c025c86c, phys: 8d47000 flags: 0 zone: c025c86c, Jun 26 16:30:57 two kernel: mem_map_t at c129ad88: phys: 9430000 flags: c zone: c025c86c, phys: 9431000 flags: c zone: c025c86c, phys: 9432000 flags: c zone: c025c86c, phys: 9433000 flags: c zone: c025c86c, Look at the flags for these two lines. The first block of mem_map_t structures (4 mem_map_t's starting at address c127bb28) each has a value for 0 in the flags field. But this second block (4 mem_map_t's starting at address c129ad88) each has a value of 0xC for the flags field. This translates to the #define PG_referenced 2 #define PG_uptodate 3 bits being set. Could someone explain to me what it means for a free page to have these bits set? -- Timur Tabi - ttabi@interactivesi.com Interactive Silicon - http://www.interactivesi.com When replying to a mailing-list message, please don't cc: me, because then I'll just get two copies of the same message. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux.eu.org/Linux-MM/