From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Wed, 23 May 2007 22:07:23 -0700 (PDT) From: Christoph Lameter Subject: Re: [patch 1/3] slob: rework freelist handling In-Reply-To: <20070524044932.GD12121@wotan.suse.de> Message-ID: References: <20070523193547.GE11115@waste.org> <20070524033925.GD14349@wotan.suse.de> <20070524041339.GC20252@wotan.suse.de> <20070524043144.GB12121@wotan.suse.de> <20070524043928.GC12121@wotan.suse.de> <20070524044932.GD12121@wotan.suse.de> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: Nick Piggin Cc: Matt Mackall , Andrew Morton , Linux Memory Management List List-ID: Booted with full slub in 16M. The additional sysfs support etc distorts the numbers a bit (300k more data) but it allows to run slabinfo. Seems that we have high kmalloc array use of 64, 128, 256, 1k, 2k and 4k. There already is a 92 and 192 sized kmalloc cache. Hmmm..... Looks like we cannot do much about this. Maybe add some odd sized slabs? A 768 byte sized? root@(none):/# mount /proc root@(none):/# mount /sys root@(none):/# free total used free shared buffers cached Mem: 11636 9976 1660 0 256 3240 -/+ buffers/cache: 6480 5156 Swap: 0 0 0 root@(none):/# slabinfo -S Name Objects Objsize Space Slabs/Part/Cpu O/S O %Fr %Ef Flg inode_cache 2471 528 1.4M 354/0/1 7 0 0 89 a :0000128 7070 128 905.2K 221/1/1 32 0 0 99 * :0001024 646 1024 663.5K 162/1/1 4 0 0 99 * :a-0000192 2730 192 536.5K 131/0/1 21 0 0 97 *a :0002048 82 2048 167.9K 41/0/1 2 0 0 100 * :0000256 528 256 139.2K 34/0/1 16 0 0 97 * :0000064 1675 64 110.5K 27/2/1 64 0 7 96 * :0004096 26 4096 106.4K 26/0/1 1 0 0 100 * radix_tree_node 148 552 90.1K 22/4/1 7 0 18 90 sighand_cache 20 2072 81.9K 20/0/1 1 0 0 50 A task_struct 20 1856 53.2K 13/6/1 2 0 46 69 :0000192 231 192 45.0K 11/0/1 21 0 0 98 * reiser_inode_cache 66 632 45.0K 11/0/1 6 0 0 92 a kmalloc-8192 5 8192 40.9K 5/0/1 1 1 0 100 :0000512 72 512 36.8K 9/0/1 8 0 0 100 * idr_layer_cache 30 528 24.5K 6/3/1 7 0 50 64 :0000032 623 32 20.4K 5/2/1 128 0 40 97 * :0000704 25 696 20.4K 5/0/1 5 0 0 84 *A :0000016 1280 16 20.4K 5/0/1 256 0 0 100 * kmalloc-8 2046 8 16.3K 4/1/1 512 0 25 99 buffer_head 111 104 12.2K 3/1/1 39 0 33 93 a vm_area_struct 48 168 8.1K 2/0/1 24 0 0 98 :0000096 84 96 8.1K 2/0/1 42 0 0 98 * sock_inode_cache 12 600 8.1K 2/0/1 6 0 0 87 Aa blkdev_requests 26 280 8.1K 2/1/1 14 0 50 88 blkdev_queue 4 1448 8.1K 2/0/1 2 0 0 70 revokefs_inode_cache 7 552 4.0K 1/0/1 7 0 0 94 Aa bdev_cache 5 720 4.0K 1/0/1 5 0 0 87 Aa mm_struct 4 792 4.0K 1/0/1 4 0 0 77 A sigqueue 25 160 4.0K 1/0/1 25 0 0 97 proc_inode_cache 7 560 4.0K 1/0/1 7 0 0 95 a Acpi-State 51 80 4.0K 1/0/1 51 0 0 99 mqueue_inode_cache 4 800 4.0K 1/0/1 4 0 0 78 A anon_vma 170 16 4.0K 1/0/1 170 0 0 66 :0000640 6 616 4.0K 1/0/1 6 0 0 90 *A shmem_inode_cache 1 712 4.0K 1/1/0 5 0 100 17 root@(none):/# slabinfo -a :0000016 <- biovec-1 kmalloc-16 :0000024 <- xfs_dabuf xfs_bmap_free_item fasync_cache swapped_entry :0000032 <- tcp_bind_bucket kmalloc-32 Acpi-Namespace :0000040 <- inotify_event_cache xfs_chashlist Acpi-Parse dnotify_cache :0000064 <- blkdev_ioc secpath_cache inet_peer_cache fs_cache uid_cache xfs_ifork Acpi-ParseExt Acpi-Operand kmalloc-64 pid biovec-4 :0000072 <- eventpoll_pwq inotify_watch_cache :0000096 <- xfs_ioend kmalloc-96 :0000128 <- kmalloc-128 request_sock_TCP sysfs_dir_cache flow_cache bio eventpoll_epi :0000192 <- xfs_ili xfs_btree_cur kmalloc-192 tw_sock_TCP :0000256 <- filp sgpool-8 mnt_cache kiocb biovec-16 kmalloc-256 arp_cache :0000320 <- ip_dst_cache kioctx xfs_buf :0000512 <- kmalloc-512 sgpool-16 :0000640 <- files_cache UNIX :0000704 <- RAW UDP-Lite UDP signal_cache :0001024 <- kmalloc-1024 biovec-64 sgpool-32 :0002048 <- kmalloc-2048 sgpool-64 biovec-128 :0004096 <- sgpool-128 biovec-256 kmalloc-4096 names_cache :a-0000192 <- skbuff_head_cache dentry -- 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-mm.org/ . Don't email: email@kvack.org