)]}' { "log": [ { "commit": "b6c24de77cbe9afe4962f0738e6444e709e83c81", "tree": "8369806f79e82bd91a80e4184972fb29f674c08c", "parents": [ "82bd3f17f4600fd374524c015826c175819f15ea" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Mon Feb 04 23:35:48 2008 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Thu Feb 21 15:27:06 2008 -0800" }, "message": "slabinfo: fall back from /sys/kernel/slab to /sys/slab\n\nI keep running upstream and mm kernels and the location of the slab\ndirectory is different since upstream still uses /sys/slab. This patch\nmakes slabinfo check /sys/slab if /sys/kernel/slab is not there. Makes\nslabinfo work on any kernel.\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "8ff12cfc009a2a38d87fa7058226fe197bb2696f", "tree": "1358ed247d3c897d8790342a978dd5078354a207", "parents": [ "1f84260c8ce3b1ce26d4c1d6dedc2f33a3a29c0c" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Thu Feb 07 17:47:41 2008 -0800" }, "committer": { "name": "Christoph Lameter", "email": "christoph@stapp.engr.sgi.com", "time": "Thu Feb 07 17:47:41 2008 -0800" }, "message": "SLUB: Support for performance statistics\n\nThe statistics provided here allow the monitoring of allocator behavior but\nat the cost of some (minimal) loss of performance. Counters are placed in\nSLUB\u0027s per cpu data structure. The per cpu structure may be extended by the\nstatistics to grow larger than one cacheline which will increase the cache\nfootprint of SLUB.\n\nThere is a compile option to enable/disable the inclusion of the runtime\nstatistics and its off by default.\n\nThe slabinfo tool is enhanced to support these statistics via two options:\n\n-D \tSwitches the line of information displayed for a slab from size\n\tmode to activity mode.\n\n-A\tSorts the slabs displayed by activity. This allows the display of\n\tthe slabs most important to the performance of a certain load.\n\n-r\tReport option will report detailed statistics on\n\nExample (tbench load):\n\nslabinfo -AD\t\t-\u003eShows the most active slabs\n\nName Objects Alloc Free %Fast\nskbuff_fclone_cache 33 111953835 111953835 99 99\n:0000192 2666 5283688 5281047 99 99\n:0001024 849 5247230 5246389 83 83\nvm_area_struct 1349 119642 118355 91 22\n:0004096 15 66753 66751 98 98\n:0000064 2067 25297 23383 98 78\ndentry 10259 28635 18464 91 45\n:0000080 11004 18950 8089 98 98\n:0000096 1703 12358 10784 99 98\n:0000128 762 10582 9875 94 18\n:0000512 184 9807 9647 95 81\n:0002048 479 9669 9195 83 65\nanon_vma 777 9461 9002 99 71\nkmalloc-8 6492 9981 5624 99 97\n:0000768 258 7174 6931 58 15\n\nSo the skbuff_fclone_cache is of highest importance for the tbench load.\nPretty high load on the 192 sized slab. Look for the aliases\n\nslabinfo -a | grep 000192\n:0000192 \u003c- xfs_btree_cur filp kmalloc-192 uid_cache tw_sock_TCP\n\trequest_sock_TCPv6 tw_sock_TCPv6 skbuff_head_cache xfs_ili\n\nLikely skbuff_head_cache.\n\n\nLooking into the statistics of the skbuff_fclone_cache is possible through\n\nslabinfo skbuff_fclone_cache\t-\u003e-r option implied if cache name is mentioned\n\n\n.... Usual output ...\n\nSlab Perf Counter Alloc Free %Al %Fr\n--------------------------------------------------\nFastpath 111953360 111946981 99 99\nSlowpath 1044 7423 0 0\nPage Alloc 272 264 0 0\nAdd partial 25 325 0 0\nRemove partial 86 264 0 0\nRemoteObj/SlabFrozen 350 4832 0 0\nTotal 111954404 111954404\n\nFlushes 49 Refill 0\nDeactivate Full\u003d325(92%) Empty\u003d0(0%) ToHead\u003d24(6%) ToTail\u003d1(0%)\n\nLooks good because the fastpath is overwhelmingly taken.\n\n\nskbuff_head_cache:\n\nSlab Perf Counter Alloc Free %Al %Fr\n--------------------------------------------------\nFastpath 5297262 5259882 99 99\nSlowpath 4477 39586 0 0\nPage Alloc 937 824 0 0\nAdd partial 0 2515 0 0\nRemove partial 1691 824 0 0\nRemoteObj/SlabFrozen 2621 9684 0 0\nTotal 5301739 5299468\n\nDeactivate Full\u003d2620(100%) Empty\u003d0(0%) ToHead\u003d0(0%) ToTail\u003d0(0%)\n\n\nDescriptions of the output:\n\nTotal:\t\tThe total number of allocation and frees that occurred for a\n\t\tslab\n\nFastpath:\tThe number of allocations/frees that used the fastpath.\n\nSlowpath:\tOther allocations\n\nPage Alloc:\tNumber of calls to the page allocator as a result of slowpath\n\t\tprocessing\n\nAdd Partial:\tNumber of slabs added to the partial list through free or\n\t\talloc (occurs during cpuslab flushes)\n\nRemove Partial:\tNumber of slabs removed from the partial list as a result of\n\t\tallocations retrieving a partial slab or by a free freeing\n\t\tthe last object of a slab.\n\nRemoteObj/Froz:\tHow many times were remotely freed object encountered when a\n\t\tslab was about to be deactivated. Frozen: How many times was\n\t\tfree able to skip list processing because the slab was in use\n\t\tas the cpuslab of another processor.\n\nFlushes:\tNumber of times the cpuslab was flushed on request\n\t\t(kmem_cache_shrink, may result from races in __slab_alloc)\n\nRefill:\t\tNumber of times we were able to refill the cpuslab from\n\t\tremotely freed objects for the same slab.\n\nDeactivate:\tStatistics how slabs were deactivated. Shows how they were\n\t\tput onto the partial list.\n\nIn general fastpath is very good. Slowpath without partial list processing is\nalso desirable. Any touching of partial list uses node specific locks which\nmay potentially cause list lock contention.\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\n\n" }, { "commit": "081248de0a0288a0ce4e1447a07ccf56aa4fae01", "tree": "2a45ac7ad26aa8f8edb9e72ac8be461be82d0271", "parents": [ "27c3a314d55b810f3b51902d1d815c714a7afbd2" ], "author": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Thu Nov 01 09:29:06 2007 -0600" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Thu Jan 24 20:40:16 2008 -0800" }, "message": "kset: move /sys/slab to /sys/kernel/slab\n\n/sys/kernel is where these things should go.\nAlso updated the documentation and tool that used this directory.\n\nCc: Kay Sievers \u003ckay.sievers@vrfy.org\u003e\nAcked-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "f32143a2fedfa299d790ca96bff9b1d2e2d6996b", "tree": "351a003cb63efe43a6e53c20483f2caeeef5737e", "parents": [ "24950898ffd161f2ea67c60f398146457bca8bf0" ], "author": { "name": "WANG Cong", "email": "xiyou.wangcong@gmail.com", "time": "Tue Oct 16 23:31:29 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Oct 17 08:43:06 2007 -0700" }, "message": "Documentation/vm/slabinfo.c: clean up this code\n\nThis patch does the following cleanups for Documentation/vm/slabinfo.c:\n\n\t- Fix two memory leaks;\n\t- Constify some char pointers;\n\t- Use snprintf instead of sprintf in case of buffer overflow;\n\t- Fix some indentations;\n\t- Other little improvements.\n\nAcked-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: WANG Cong \u003cxiyou.wangcong@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ac07860264bd2b18834d3fa3be47032115524cea", "tree": "f0456c2ccbbf26a69cd0a137132e8e141ed40780", "parents": [ "1ceef40249f21eceabf8633934d94962e7d8e1d7" ], "author": { "name": "Jesper Juhl", "email": "jesper.juhl@gmail.com", "time": "Wed Aug 08 16:31:30 2007 -0700" }, "committer": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Thu Aug 09 21:57:16 2007 -0700" }, "message": "SLUB: Fix format specifier in Documentation/vm/slabinfo.c\n\nThere\u0027s a little problem in Documentation/vm/slabinfo.c\nThe code is using \"%d\" in a printf() call to print an \u0027unsigned long\u0027.\nThis patch corrects it to use \"%lu\" instead.\n\nSigned-off-by: Jesper Juhl \u003cjesper.juhl@gmail.com\u003e\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\n" }, { "commit": "32f9306b161a5e0c892132f7227f36de94090289", "tree": "69a04fdd1d38e535af7729be18e4244dad4b1013", "parents": [ "223a10a98135da38d3668973d72cdffb1ced4b7d" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Fri May 18 00:36:43 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri May 18 08:17:17 2007 -0700" }, "message": "slub: another slabinfo fix\n\nThe slab manipulation functions should not be triggered by slabs that\nare unresovable in the subset of slabs selected on the command line.\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "eefaca9c3246f3daf56e7ed02987f79abcee7087", "tree": "1eac21979f536f7a08ae77d4b59f1c615eef5b94", "parents": [ "ade3aff25fb2dce76e2a9b53e1334bd0a174f739" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Wed May 16 22:10:55 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu May 17 05:23:03 2007 -0700" }, "message": "SLUB: slabinfo fixes\n\nAlign the output of % with K/M/G of sizes.\n\nCheck for empty NUMA information to avoid segfault on !NUMA.\n\n-r should work directly not only if we match a single slab\n without additional options.\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a87615b8f9e2349f6d3770af3d72fd6a41ab4239", "tree": "5e2608dad1ad746b0552ef52afec9298320045ac", "parents": [ "be7b3fbcef34452127bed93632b8e788f685d70e" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Wed May 09 02:32:37 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed May 09 12:30:44 2007 -0700" }, "message": "SLUB: slabinfo upgrade\n\n-e Show empty slabs\n-d Modification of slab debug options at runtime\n-o Operations. Display of ctor / dtor etc.\n-r Report: Display all available information about a slabcache.\n\nCleanup tracking display and make it work right.\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c09d87517298fd01543739ba26987645deb4e6a9", "tree": "dddbcc938e28787ea1c411e2cc918411ffaae93a", "parents": [ "352434211dad370316155d90d7dab590519f465b" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Sun May 06 14:49:48 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Mon May 07 12:12:54 2007 -0700" }, "message": "slub: add slabinfo tool\n\nAdd the tool which gets reports about slabs to the VM documentation directory.\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" } ] }