)]}' { "log": [ { "commit": "60a3cdd0639473c79c253bc08c8ef8f882cca107", "tree": "e72a5ca87873b286b02966a390277f828139d350", "parents": [ "b1721d0da266b4af8cb4419473b4ca36206ab200" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Mar 03 12:38:52 2008 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Sat Apr 26 17:44:55 2008 +0200" }, "message": "x86: add optimized inlining\n\nadd CONFIG_OPTIMIZE_INLINING\u003dy.\n\nallow gcc to optimize the kernel image\u0027s size by uninlining\nfunctions that have been marked \u0027inline\u0027. Previously gcc was\nforced by Linux to always-inline these functions via a gcc\nattribute:\n\n #define inline\tinline __attribute__((always_inline))\n\nEspecially when the user has already selected\nCONFIG_OPTIMIZE_FOR_SIZE\u003dy this can make a huge difference in\nkernel image size (using a standard Fedora .config):\n\n text data bss dec hex filename\n 5613924 562708 3854336 10030968 990f78 vmlinux.before\n 5486689 562708 3854336 9903733 971e75 vmlinux.after\n\nthat\u0027s a 2.3% text size reduction (!).\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "185c045c245f46485ad8bbd8cc1100e986ff3f13", "tree": "fc08cd474bf6c7c9b0c077a069255d04b8804fe4", "parents": [ "bfc734b24671b2639218ae2ef53af91dfd30b6c9" ], "author": { "name": "Harvey Harrison", "email": "harvey.harrison@gmail.com", "time": "Sat Feb 09 23:24:09 2008 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 09 23:24:09 2008 +0100" }, "message": "x86, core: remove CONFIG_FORCED_INLINING\n\nOther than the defconfigs, remove the entry in compiler-gcc4.h,\nKconfig.debug and feature-removal-schedule.txt.\n\nSigned-off-by: Harvey Harrison \u003charvey.harrison@gmail.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "a9a832902fccc958caa339677dbff766e51b02c3", "tree": "a74ffce96d614210e40317bc035b71ce64293374", "parents": [ "a6b68076fd5d7de04e6847ed9d7d212e7dd6beb5" ], "author": { "name": "Andi Kleen", "email": "ak@suse.de", "time": "Wed Jan 30 13:32:49 2008 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Wed Jan 30 13:32:49 2008 +0100" }, "message": "x86: Set CFQ as default in 32-bit defconfig\n\nSomeone complained that the 32-bit defconfig contains AS as default IO\nscheduler. Change that to CFQ.\n\nSigned-off-by: Andi Kleen \u003cak@suse.de\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "b263295dbffd33b0fbff670720fa178c30e3392a", "tree": "014b402cb65b17e1521ed356e5c225083ea578dd", "parents": [ "4ebd1290ba12121d66285cc06987ca97bcdfc55b" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Wed Jan 30 13:30:47 2008 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Wed Jan 30 13:30:47 2008 +0100" }, "message": "x86: 64-bit, make sparsemem vmemmap the only memory model\n\nUse sparsemem as the only memory model for UP, SMP and NUMA. Measurements\nindicate that DISCONTIGMEM has a higher overhead than sparsemem. And\nFLATMEMs benefits are minimal. So I think its best to simply standardize\non sparsemem.\n\nResults of page allocator tests (test can be had via git from slab git\ntree branch tests)\n\nMeasurements in cycle counts. 1000 allocations were performed and then the\naverage cycle count was calculated.\n\nOrder\tFlatMem\tDiscontig\tSparseMem\n0\t 639\t 665\t\t 641\n1\t 567\t 647\t\t 593\n2\t 679\t 774\t\t 692\n3\t 763\t 967\t\t 781\n4\t 961\t 1501\t\t 962\n5\t 1356\t 2344\t\t 1392\n6\t 2224\t 3982\t\t 2336\n7\t 4869\t 7225\t\t 5074\n8\t12500\t14048\t\t12732\n9\t27926\t28223\t\t28165\n10\t58578\t58714\t\t58682\n\n(Note that FlatMem is an SMP config and the rest NUMA configurations)\n\nMemory use:\n\nSMP Sparsemem\n-------------\n\nKernel size:\n\n text data bss dec hex filename\n3849268 397739 1264856 5511863 541ab7 vmlinux\n\n total used free shared buffers cached\nMem: 8242252 41164 8201088 0 352 11512\n-/+ buffers/cache: 29300 8212952\nSwap: 9775512 0 9775512\n\nSMP Flatmem\n-----------\n\nKernel size:\n\n text data bss dec hex filename\n3844612 397739 1264536 5506887 540747 vmlinux\n\nSo 4.5k growth in text size vs. FLATMEM.\n\n total used free shared buffers cached\nMem: 8244052 40544 8203508 0 352 11484\n-/+ buffers/cache: 28708 8215344\n\n2k growth in overall memory use after boot.\n\nNUMA discontig:\n\n text data bss dec hex filename\n3888124 470659 1276504 5635287 55fcd7 vmlinux\n\n total used free shared buffers cached\nMem: 8256256 56908 8199348 0 352 11496\n-/+ buffers/cache: 45060 8211196\nSwap: 9775512 0 9775512\n\nNUMA sparse:\n\n text data bss dec hex filename\n3896428 470659 1276824 5643911 561e87 vmlinux\n\n8k text growth. Given that we fully inline virt_to_page and friends now\nthat is rather good.\n\n total used free shared buffers cached\nMem: 8264720 57240 8207480 0 352 11516\n-/+ buffers/cache: 45372 8219348\nSwap: 9775512 0 9775512\n\nThe total available memory is increased by 8k.\n\nThis patch makes sparsemem the default and removes discontig and\nflatmem support from x86.\n\n[ akpm@linux-foundation.org: allnoconfig build fix ]\n\nAcked-by: Andi Kleen \u003cak@suse.de\u003e\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "966396d3a05c8049fce5c81c49138e5ee1b05443", "tree": "11fff010a4aecd965bed52719d55d8fa4b40b055", "parents": [ "395624fcddd178de01a78aa88670a86ec919de77" ], "author": { "name": "Joerg Roedel", "email": "joerg.roedel@amd.com", "time": "Wed Oct 24 12:49:48 2007 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Oct 30 00:22:22 2007 +0100" }, "message": "x86 gart: rename CONFIG_IOMMU to CONFIG_GART_IOMMU\n\nThis patch renames the IOMMU config option to GART_IOMMU because in fact it\nmeans the GART and not general support for an IOMMU on x86.\n\nSigned-off-by: Joerg Roedel \u003cjoerg.roedel@amd.com\u003e\nAcked-by: Muli Ben-Yehuda \u003cmuli@il.ibm.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "2266cfd50de3872e877eeca3dd4a6f940f22ba60", "tree": "362afb85ef2d13b12c479b3183e92aacb5da64db", "parents": [ "74b469f2e6b1589118b6ac05c7256add01e572d2" ], "author": { "name": "Sam Ravnborg", "email": "sam@ravnborg.org", "time": "Thu Oct 25 20:31:19 2007 +0200" }, "committer": { "name": "Sam Ravnborg", "email": "sam@ravnborg.org", "time": "Thu Oct 25 22:27:41 2007 +0200" }, "message": "x86: move defconfig files for i386 and x86_64 to x86\n\nWith some small changes to kconfig makefile we can now\nlocate the defconfig files for i386 and x86_64 in\nthe configs/ subdirectory under x86.\nmake ARCH\u003di386 defconfig and make defconfig\nworks as expected also after this change.\nBut arch maintainers shall now update a defconfig file in\nthe configs/ directory.\n\nSigned-off-by: Sam Ravnborg \u003csam@ravnborg.org\u003e\n" } ] }