)]}' { "log": [ { "commit": "8b56a30caaf9bc1850784f196636c5f550cc7577", "tree": "0e1be3d4f929a385c666fd7b269763f59fe215e5", "parents": [ "6abe4a87f7bc7978705c386dbba0ca0c7790b3ec" ], "author": { "name": "Sachin Bhamare", "email": "sbhamare@panasas.com", "time": "Wed Mar 14 18:01:45 2012 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon May 21 12:24:01 2012 +0300" }, "message": "exofs: Add SYSFS info for autologin/pNFS export\n\nIntroduce sysfs infrastructure for exofs cluster filesystem.\n\nEach OSD target shows up as below in the sysfs hierarchy:\n\t/sys/fs/exofs/\u003cosdname\u003e_\u003cpartition_id\u003e/devX\n\nWhere \u003cosdname\u003e_\u003cpartition_id\u003e is the unique identification\nof a Superblock.\n\nWhere devX: 0 \u003c\u003d X \u003c device_table_size. They are ordered\nin device-table order as specified to the mkfs.exofs command\n\nEach OSD device devX has following attributes :\n\tosdname - ReadOnly\n\tsystemid - ReadOnly\n\turi - Read/Write\n\nIt is up to user-mode to update devX/uri for support of\nautologin.\n\nThese sysfs information are used both for autologin as well\nas support for exporting exofs via a pNFSD server in user-mode.\n(.eg NFS-Ganesha)\n\nSigned-off-by: Sachin Bhamare \u003csbhamare@panasas.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "6abe4a87f7bc7978705c386dbba0ca0c7790b3ec", "tree": "02f99246491cc2a76bdb592860537c63ee0ef480", "parents": [ "36be50515fe2aef61533b516fa2576a2c7fe7664" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed May 16 14:22:21 2012 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun May 20 19:42:41 2012 +0300" }, "message": "exofs: Fix CRASH on very early IO errors.\n\nIf at exofs_fill_super() we had an early termination\ndo to any error, like an IO error while reading the\nsuper-block. We would crash inside exofs_free_sbi().\n\nThis is because sbi-\u003eoc.numdevs was set to 1, before\nwe actually have a device table at all.\n\nFix it by moving the sbi-\u003eoc.numdevs \u003d 1 to after the\nallocation of the device table.\n\nReported-by: Johannes Schild \u003cJSchild@gmx.de\u003e\n\nStable: This is a bug since v3.2.0\nCC: Stable Tree \u003cstable@kernel.org\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "afb9bd704c7116076879352a2cc2c43aa12c1e14", "tree": "b622fba5044ef3111050ca9dacce207ba49fea39", "parents": [ "58df9b387c5f240566ac15b7fa5136f5a35bb19a", "72749a270b6d254b4a018e290b853c27edb2fa62" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Mar 28 20:04:27 2012 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Mar 28 20:04:27 2012 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.open-osd.org/linux-open-osd\n\nPull trivial exofs changes from Boaz Harrosh:\n \"Just nothingness really. The big exofs changes are reserved for the\n next merge window.\"\n\n* \u0027for-linus\u0027 of git://git.open-osd.org/linux-open-osd:\n exofs: Cap on the memcpy() size\n exofs: (trivial) Fix typo in super.c\n exofs: fix endian conversion in exofs_sync_fs()\n" }, { "commit": "48fde701aff662559b38d9a609574068f22d00fe", "tree": "aa6b203dc671b51d58575b65eb08310ff8309b60", "parents": [ "6b4231e2f92adbcf96fb2a3fa751d7ca0a61b21f" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Jan 08 22:15:13 2012 -0500" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Tue Mar 20 21:29:35 2012 -0400" }, "message": "switch open-coded instances of d_make_root() to new helper\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "8de52778798fe39660a8d6b26f290e0c93202761", "tree": "56384beb7863c1f338f2b66b509bf58eea5a99c2", "parents": [ "c16fa4f2ad19908a47c63d8fa436a1178438c7e7" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Feb 06 12:45:27 2012 -0500" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Tue Mar 20 21:29:32 2012 -0400" }, "message": "vfs: check i_nlink limits in vfs_{mkdir,rename_dir,link}\n\nNew field of struct super_block - -\u003es_max_links. Maximal allowed\nvalue of -\u003ei_nlink or 0; in the latter case all checks still need\nto be done in -\u003elink/-\u003emkdir/-\u003erename instances. Note that this\nlimit applies both to directoris and to non-directories.\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "72749a270b6d254b4a018e290b853c27edb2fa62", "tree": "cde350c359d5dc7328e6a96f867ca6d671b4bde8", "parents": [ "3e57638bb1469ba2705456e9fd4063d1890341e1" ], "author": { "name": "Dan Carpenter", "email": "dan.carpenter@oracle.com", "time": "Mon Jan 30 10:59:49 2012 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Mar 19 13:39:12 2012 -0700" }, "message": "exofs: Cap on the memcpy() size\n\nThis data comes from the device, so probably it\u0027s fairly trustworthy but\nit makes the static checkers happy if we check it.\n\n[Boaz]\n the system_id_len is zero, if not present, or always OSD_SYSTEMID_LEN.\n So always copy OSD_SYSTEMID_LEN bytes.\n\nSigned-off-by: Dan Carpenter \u003cdan.carpenter@oracle.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "3e57638bb1469ba2705456e9fd4063d1890341e1", "tree": "92a2bf22980abb72a251d836ee03022b96049db9", "parents": [ "b6d1f2dd61de7e696c75fbf39fb99cf41a189084" ], "author": { "name": "Masanari Iida", "email": "standby24x7@gmail.com", "time": "Wed Feb 01 23:53:46 2012 +0900" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Mar 19 13:39:12 2012 -0700" }, "message": "exofs: (trivial) Fix typo in super.c\n\nCorrect spelling \"faild\" to \"failed\" in\nfs/exofs/super.c\n\nSigned-off-by: Masanari Iida \u003cstandby24x7@gmail.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "b6d1f2dd61de7e696c75fbf39fb99cf41a189084", "tree": "4e4304318f3b5a0a404ea2b0a7539c8f97689819", "parents": [ "192cfd58774b4d17b2fe8bdc77d89c2ef4e0591d" ], "author": { "name": "Dan Carpenter", "email": "dan.carpenter@oracle.com", "time": "Mon Nov 21 17:43:11 2011 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Mar 19 13:39:11 2012 -0700" }, "message": "exofs: fix endian conversion in exofs_sync_fs()\n\nfscb-\u003es_numfiles is an __le64 field so we need to use cpu_to_le64()\nto get a little endian 64 bit on big endian systems.\n\nSigned-off-by: Dan Carpenter \u003cdan.carpenter@oracle.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "da01636a6511c3bd0c1cf546c47b8e92a837a613", "tree": "394f40f4763d444ace4990ff1a5d21b386f83167", "parents": [ "3850aba74873aa47fefe6900b99f42f5e656a6e7" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Jan 08 19:45:28 2012 -0500" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Jan 08 20:19:12 2012 -0500" }, "message": "exofs: oops after late failure in mount\n\nWe have already set -\u003es_root, so -\u003eput_super() is going to be called.\nFreeing -\u003es_fs_info is a bloody bad idea when it\u0027s going to be\ndereferenced very shortly...\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "6b520e0565422966cdf1c3759bd73df77b0f248c", "tree": "f63a26afa7342eb59b125271b16e30a219b59094", "parents": [ "2a79f17e4a641a2f463cb512cb0ec349844a147b" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Dec 12 15:51:45 2011 -0500" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Tue Jan 03 22:52:40 2012 -0500" }, "message": "vfs: fix the stupidity with i_dentry in inode destructors\n\nSeeing that just about every destructor got that INIT_LIST_HEAD() copied into\nit, there is no point whatsoever keeping this INIT_LIST_HEAD in inode_init_once();\nthe cost of taking it into inode_init_always() will be negligible for pipes\nand sockets and negative for everything else. Not to mention the removal of\nboilerplate code from -\u003edestroy_inode() instances...\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "143cb494cb6662e37c4020b7fe9839837f718e56", "tree": "249abcf5afe251a9244a414f9d66c73b9458274e", "parents": [ "afeacc8c1f38b7bb93d4bc7b4ba04c2605061ef0" ], "author": { "name": "Paul Gortmaker", "email": "paul.gortmaker@windriver.com", "time": "Fri Jul 01 14:23:34 2011 -0400" }, "committer": { "name": "Paul Gortmaker", "email": "paul.gortmaker@windriver.com", "time": "Mon Oct 31 19:30:31 2011 -0400" }, "message": "fs: add module.h to files that were implicitly using it\n\nSome files were using the complete module.h infrastructure without\nactually including the header at all. Fix them up in advance so\nonce the implicit presence is removed, we won\u0027t get failures like this:\n\n CC [M] fs/nfsd/nfssvc.o\nfs/nfsd/nfssvc.c: In function \u0027nfsd_create_serv\u0027:\nfs/nfsd/nfssvc.c:335: error: \u0027THIS_MODULE\u0027 undeclared (first use in this function)\nfs/nfsd/nfssvc.c:335: error: (Each undeclared identifier is reported only once\nfs/nfsd/nfssvc.c:335: error: for each function it appears in.)\nfs/nfsd/nfssvc.c: In function \u0027nfsd\u0027:\nfs/nfsd/nfssvc.c:555: error: implicit declaration of function \u0027module_put_and_exit\u0027\nmake[3]: *** [fs/nfsd/nfssvc.o] Error 1\n\nSigned-off-by: Paul Gortmaker \u003cpaul.gortmaker@windriver.com\u003e\n" }, { "commit": "5a51c0c7e9a913649aa65d8233470682bcbb7694", "tree": "4ed43b3a2f9fade227e098771a5bed66f6a71dc4", "parents": [ "3bd9856857339d7ee8c4ad50030583f1b9415c39" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Sep 28 13:18:45 2011 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Fri Oct 14 18:54:41 2011 +0200" }, "message": "ore/exofs: Define new ore_verify_layout\n\nAll users of the ore will need to check if current code\nsupports the given layout. For example RAID5/6 is not\ncurrently supported.\n\nSo move all the checks from exofs/super.c to a new\nore_verify_layout() to be used by ore users.\n\nNote that any new layout should be passed through the\nore_verify_layout() because the ore engine will prepare\nand verify some internal members of ore_layout, and\nassumes it\u0027s called.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "d866d875f68fdeae63df334d291fe138dc636d96", "tree": "9606674db2311ab869640526ef245aaa7fbf4ea8", "parents": [ "eb507bc18969f63b8968034144fd69706c492516" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Sep 28 14:43:09 2011 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Oct 04 12:13:59 2011 +0200" }, "message": "ore/exofs: Change the type of the devices array (API change)\n\nIn the pNFS obj-LD the device table at the layout level needs\nto point to a device_cache node, where it is possible and likely\nthat many layouts will point to the same device-nodes.\n\nIn Exofs we have a more orderly structure where we have a single\narray of devices that repeats twice for a round-robin view of the\ndevice table\n\nThis patch moves to a model that can be used by the pNFS obj-LD\nwhere struct ore_components holds an array of ore_dev-pointers.\n(ore_dev is newly defined and contains a struct osd_dev *od\n member)\n\nEach pointer in the array of pointers will point to a bigger\nuser-defined dev_struct. That can be accessed by use of the\ncontainer_of macro.\n\nIn Exofs an __alloc_dev_table() function allocates the\nore_dev-pointers array as well as an exofs_dev array, in one\nallocation and does the addresses dance to set everything pointing\ncorrectly. It still keeps the double allocation trick for the\ninodes round-robin view of the table.\n\nThe device table is always allocated dynamically, also for the\nsingle device case. So it is unconditionally freed at umount.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "8d2d83a8352b0f9c1da82c36f741722f2960feea", "tree": "4a96539dce29c654def762d502b41da8de7cbce7", "parents": [ "5bf696dad4beecb6174e701c97e1f2574e6a2c96" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Aug 10 14:15:02 2011 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Oct 03 17:07:51 2011 +0200" }, "message": "exofs: Remove unused data_map member from exofs_sb_info\n\nThe struct pnfs_osd_data_map data_map member of exofs_sb_info was\nnever used after mount. In fact all it\u0027s members were duplicated\nby the ore_layout structure. So just remove the duplicated information.\n\nAlso removed some stupid, but perfectly supported, restrictions on\nlayout parameters. The case where num_devices is not divisible by\nmirror_count+1 is perfectly fine since the rotating device view\nwill eventually use all the devices it can get.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\nSigned-off-by: Benny Halevy \u003cbhalevy@tonian.com\u003e\n" }, { "commit": "5bf696dad4beecb6174e701c97e1f2574e6a2c96", "tree": "b4cfd7e00a61310aaf5ac90e670c840184fc115d", "parents": [ "de74b05ace743b4a7aefad9e9b33ff899979b34a" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Sep 28 11:39:59 2011 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Oct 03 17:07:50 2011 +0200" }, "message": "exofs: Rename struct ore_components comps \u003d\u003e oc\n\nore_components already has a comps member so this leads\nto things like comps-\u003ecomps which is annoying. the name oc\nwas already used in new code. So rename all old usage of\nore_components comps \u003d\u003e ore_components oc.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "de74b05ace743b4a7aefad9e9b33ff899979b34a", "tree": "9f62853e67803cf406eaf35238a429128f2272d3", "parents": [ "1958c7c284298f0d17a1ab3820c77c40ea3d3711" ], "author": { "name": "H Hartley Sweeten", "email": "hartleys@visionengravers.com", "time": "Fri Sep 23 13:46:51 2011 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Oct 03 17:07:29 2011 +0200" }, "message": "exofs/super.c: local functions should be static\n\nThis quiets the following sparse noise:\n\nwarning: symbol \u0027exofs_sync_fs\u0027 was not declared. Should it be static?\nwarning: symbol \u0027exofs_free_sbi\u0027 was not declared. Should it be static?\nwarning: symbol \u0027exofs_get_parent\u0027 was not declared. Should it be static?\n\nSigned-off-by: H Hartley Sweeten \u003chsweeten@visionengravers.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "8ff660ab85f524bdc7652eb5d38aaef1d66aa9c7", "tree": "c4a29cde4fc8654ae00e65cb520e13f9fe7f4e08", "parents": [ "9e9db45649eb5d3ee5622fdad741914ecf1016a0" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sat Aug 06 19:26:31 2011 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sat Aug 06 19:36:18 2011 -0700" }, "message": "exofs: Rename raid engine from exofs/ios.c \u003d\u003e ore\n\nORE stands for \"Objects Raid Engine\"\n\nThis patch is a mechanical rename of everything that was in ios.c\nand its API declaration to an ore.c and an osd_ore.h header. The ore\nengine will later be used by the pnfs objects layout driver.\n\n* File ios.c \u003d\u003e ore.c\n\n* Declaration of types and API are moved from exofs.h to a new\n osd_ore.h\n\n* All used types are prefixed by ore_ from their exofs_ name.\n\n* Shift includes from exofs.h to osd_ore.h so osd_ore.h is\n independent, include it from exofs.h.\n\nOther than a pure rename there are no other changes. Next patch\nwill move the ore into it\u0027s own module and will export the API\nto be used by exofs and later the layout driver\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "9e9db45649eb5d3ee5622fdad741914ecf1016a0", "tree": "19ab9e1431e3d6535cef3f2cba6fcff12bb6ba6c", "parents": [ "85e44df4748670a1a7d8441b2d75843cdebc478a" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Fri Aug 05 15:06:04 2011 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sat Aug 06 19:35:32 2011 -0700" }, "message": "exofs: ios: Move to a per inode components \u0026 device-table\n\nExofs raid engine was saving on memory space by having a single layout-info,\nsingle pid, and a single device-table, global to the filesystem. Then passing\na credential and object_id info at the io_state level, private for each\ninode. It would also devise this contraption of rotating the device table\nview for each inode-\u003eino to spread out the device usage.\n\nThis is not compatible with the pnfs-objects standard, demanding that\neach inode can have it\u0027s own layout-info, device-table, and each object\ncomponent it\u0027s own pid, oid and creds.\n\nSo: Bring exofs raid engine to be usable for generic pnfs-objects use by:\n\n* Define an exofs_comp structure that holds obj_id and credential info.\n\n* Break up exofs_layout struct to an exofs_components structure that holds a\n possible array of exofs_comp and the array of devices + the size of the\n arrays.\n\n* Add a \"comps\" parameter to get_io_state() that specifies the ids creds\n and device array to use for each IO.\n\n This enables to keep the layout global, but the device-table view, creds\n and IDs at the inode level. It only adds two 64bit to each inode, since\n some of these members already existed in another form.\n\n* ios raid engine now access layout-info and comps-info through the passed\n pointers. Everything is pre-prepared by caller for generic access of\n these structures and arrays.\n\nAt the exofs Level:\n\n* Super block holds an exofs_components struct that holds the device\n array, previously in layout. The devices there are in device-table\n order. The device-array is twice bigger and repeats the device-table\n twice so now each inode\u0027s device array can point to a random device\n and have a round-robin view of the table, making it compatible to\n previous exofs versions.\n\n* Each inode has an exofs_components struct that is initialized at\n load time, with it\u0027s own view of the device table IDs and creds.\n When doing IO this gets passed to the io_state together with the\n layout.\n\nWhile preforming this change. Bugs where found where credentials with the\nwrong IDs where used to access the different SB objects (super.c). As well\nas some dead code. It was never noticed because the target we use does not\ncheck the credentials.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "85e44df4748670a1a7d8441b2d75843cdebc478a", "tree": "c5bc0cdf7dad56cc6f3a38f99c88f62325a1e029", "parents": [ "e1042ba0991aab80ced34f7dade6ec25f22b4304" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon May 16 15:26:47 2011 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sat Aug 06 19:35:31 2011 -0700" }, "message": "exofs: Move exofs specific osd operations out of ios.c\n\nios.c will be moving to an external library, for use by the\nobjects-layout-driver. Remove from it some exofs specific functions.\n\nAlso g_attr_logical_length is used both by inode.c and ios.c\nmove definition to the later, to keep it independent\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "9ce730475e1b950d78a69c1be3410109c103ac98", "tree": "7bc785ad8990932b168a51036e39cadbb03bddfa", "parents": [ "6d4073e88132259485ef1b2c88daa5e50c95789c" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Aug 03 20:18:01 2011 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Aug 04 12:35:23 2011 -0700" }, "message": "exofs: Small cleanup of exofs_fill_super\n\nSmall cleanup that unifies duplicated code used in both the\nerror and success cases\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "6d4073e88132259485ef1b2c88daa5e50c95789c", "tree": "04876cbe47e462bca7dad51ec33ccb29080d5680", "parents": [ "26ae93c2dc7152463d319c28768f242a11a54620" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Jul 27 17:51:53 2011 -0700" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Aug 04 12:35:20 2011 -0700" }, "message": "exofs: BUG: Avoid sbi realloc\n\nSince the beginning we realloced the sbi structure when a bigger\nthen one device table was specified. (I know that was really stupid).\n\nThen much later when \"register bdi\" was added (By Jens) it was\nregistering the pointer to sbi-\u003ebdi before the realloc.\n\nWe never saw this problem because up till now the realloc did not\ndo anything since the device table was small enough to fit in the\noriginal allocation. But once we starting testing with large device\ntables (Bigger then 28) we noticed the crash of writeback operating\non a deallocated pointer.\n\n* Avoid the all mess by allocating the device-table as a second array\n and get rid of the variable-sized structure and the rest of this\n mess.\n* Take the chance to clean near by structures and comments.\n* Add a needed dprint on startup to indicate the loaded layout.\n* Also move the bdi registration to the very end because it will\n only fail in a low memory, which will probably fail before hand.\n There are many more likely causes to not load before that. This\n way the error handling is made simpler. (Just doing this would be\n enough to fix the BUG)\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "a803b8067e317832d6a251c5b0486e36a4f81922", "tree": "db7a40edece0e43f870bc33776bb1a7e4fb13006", "parents": [ "1b71fe2efa31cd18c865db474a4cd473b6ab5281" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Fri Jul 08 20:56:55 2011 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Jul 17 23:20:29 2011 -0400" }, "message": "fix exofs -\u003eget_parent()\n\nNULL is not a possible return value for that method, TYVM...\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "a49fb4c3d035ab516507b31ec3bd49263caee14d", "tree": "aa44d448190dacad75b30fe86c49a72ad9cecc0d", "parents": [ "1cea312ad49d9cb964179a784fedb1fcfe396283" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Feb 07 18:12:15 2011 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 15 15:02:52 2011 +0200" }, "message": "exofs: deprecate the commands pending counter\n\nOne leftover from the days of IBM\u0027s original code, is an SB counter\nthat counts in-flight asynchronous commands. And a piece of code that\nwaits for the counter to reach zero at unmount. I guess it might have\nbeen needed then, cause of some reference missing or something.\n\nI\u0027m not removing it yet but am putting a warning message if ever this\ncounter triggers at unmount. If I\u0027ll never see it triggers or reported\nI\u0027ll remove the counter for good.\n(I had this print as a debug output for a long time and never had it\n trigger)\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "1cea312ad49d9cb964179a784fedb1fcfe396283", "tree": "27c45af006b48b1a079698605ea9007398f652b5", "parents": [ "9ed96484311b89360b80a4181d856cbdb21630fd" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Feb 03 17:53:25 2011 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 15 15:02:51 2011 +0200" }, "message": "exofs: Write sbi-\u003es_nextid as part of the Create command\n\nBefore when creating a new inode, we\u0027d set the sb-\u003es_dirt flag,\nand sometime later the system would write out s_nextid as part\nof the sb_info. Also on inode sync we would force the sb sync\nas well.\n\nDefine the s_nextid as a new partition attribute and set it\nevery time we create a new object.\nAt mount we read it from it\u0027s new place.\n\nWe now never set sb-\u003es_dirt anywhere in exofs. write_super\nis actually never called. The call to exofs_write_super from\nexofs_put_super is also removed because the VFS always calls\n-\u003esync_fs before calling -\u003eput_super twice.\n\nTo stay backward-and-forward compatible we also write the old\ns_nextid in the super_block object at unmount, and support zero\nlength attribute on mount.\n\nThis also fixes a BUG where in layouts when group_width was not\na divisor of EXOFS_SUPER_ID (0x10000) the s_nextid was not read\nfrom the device it was written to. Because of the sliding window\nlayout trick, and because the read was always done from the 0\ndevice but the write was done via the raid engine that might slide\nthe device view. Now we read and write through the raid engine.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "9ed96484311b89360b80a4181d856cbdb21630fd", "tree": "ef372f43e6bfdf86d273d80491e7de801b69395f", "parents": [ "66cd6cad4919f980dd21307d0150ff251762a264" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Jan 31 14:32:14 2011 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 15 15:02:51 2011 +0200" }, "message": "exofs: Add option to mount by osdname\n\nIf /dev/osd* devices are shuffled because more devices\nwhere added, and/or login order has changed. It is hard to\nmount the FS you want.\n\nAdd an option to mount by osdname. osdname is any osd-device\u0027s\nosdname as specified to the mkfs.exofs command when formatting\nthe osd-devices.\nThe new mount format is:\n\tOPT\u003d\"osdname\u003d$UUID0,pid\u003d$PID,_netdev\"\n\tmount -t exofs -o $OPT $DEV_OSD0 $MOUNTDIR\n\nif \"osdname\u003d\" is specified in options above $DEV_OSD0 is\nignored and can be empty.\n\nAlso while at it: Removed some old unused Opt_* enums.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "66cd6cad4919f980dd21307d0150ff251762a264", "tree": "4f2d9dc0f9070ce829d03097542e7968e02df41a", "parents": [ "97178b7b6c84bd14660b89474d27931a1ea65c66" ], "author": { "name": "bharrosh@panasas.com", "email": "bharrosh@panasas.com", "time": "Thu Oct 07 14:28:18 2010 -0400" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 15 15:02:50 2011 +0200" }, "message": "exofs: Override read-ahead to align on stripe_size\n\n* Set all inode-\u003ei_mapping-\u003ebacking_dev_info to point to\n the per super-block sb-\u003es_bdi.\n\n* Calculating a read_ahead that is:\n - preferable 2 stripes long\n (Future patch will add a mount option to override this)\n - Minimum 128K aligned up to stripe-size\n - Caped to maximum-IO-sizes round down to stripe_size.\n (Max sizes are governed by max bio-size that fits in a page\n times number-of-devices)\n\nCC: Marc Dionne \u003cmarc.c.dionne@gmail.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "2c722c9a47d1369e2685b85288e78c469a081238", "tree": "2acb65a6c2ba1d756b317e6ab7ee9624851b83ed", "parents": [ "521cb40b0c44418a4fd36dc633f575813d59a43d" ], "author": { "name": "Tobias Klauser", "email": "tklauser@distanz.ch", "time": "Thu Dec 09 15:55:21 2010 +0100" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 15 12:33:42 2011 +0200" }, "message": "exofs: Remove redundant unlikely()\n\nIS_ERR() already implies unlikely(), so it can be omitted here.\n\nSigned-off-by: Tobias Klauser \u003ctklauser@distanz.ch\u003e\n" }, { "commit": "fa0d7e3de6d6fc5004ad9dea0dd6b286af8f03e9", "tree": "203e0f73883e4c26b5597e36042386a1237dab35", "parents": [ "77812a1ef139d84270d27faacc0630c887411013" ], "author": { "name": "Nick Piggin", "email": "npiggin@kernel.dk", "time": "Fri Jan 07 17:49:49 2011 +1100" }, "committer": { "name": "Nick Piggin", "email": "npiggin@kernel.dk", "time": "Fri Jan 07 17:50:26 2011 +1100" }, "message": "fs: icache RCU free inodes\n\nRCU free the struct inode. This will allow:\n\n- Subsequent store-free path walking patch. The inode must be consulted for\n permissions when walking, so an RCU inode reference is a must.\n- sb_inode_list_lock to be moved inside i_lock because sb list walkers who want\n to take i_lock no longer need to take sb_inode_list_lock to walk the list in\n the first place. This will simplify and optimize locking.\n- Could remove some nested trylock loops in dcache code\n- Could potentially simplify things a bit in VM land. Do not need to take the\n page lock to follow page-\u003emapping.\n\nThe downsides of this is the performance cost of using RCU. In a simple\ncreat/unlink microbenchmark, performance drops by about 10% due to inability to\nreuse cache-hot slab objects. As iterations increase and RCU freeing starts\nkicking over, this increases to about 20%.\n\nIn cases where inode lifetimes are longer (ie. many inodes may be allocated\nduring the average life span of a single inode), a lot of this cache reuse is\nnot applicable, so the regression caused by this patch is smaller.\n\nThe cache-hot regression could largely be avoided by using SLAB_DESTROY_BY_RCU,\nhowever this adds some complexity to list walking and store-free path walking,\nso I prefer to implement this at a later date, if it is shown to be a win in\nreal situations. I haven\u0027t found a regression in any non-micro benchmark so I\ndoubt it will be a problem.\n\nSigned-off-by: Nick Piggin \u003cnpiggin@kernel.dk\u003e\n" }, { "commit": "3c26ff6e499ee7e6f9f2bc7da5f2f30d80862ecf", "tree": "bd758d7f15f24aed225a64de77cc535785c50f96", "parents": [ "fc14f2fef682df677d64a145256dbd263df2aa7b" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Sun Jul 25 11:46:36 2010 +0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Fri Oct 29 04:16:31 2010 -0400" }, "message": "convert get_sb_nodev() users\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "bf25db365428dbd182768baa9850bef7afaac80d", "tree": "4373037b00459643f0d661672eb6688f80800d72", "parents": [ "682c30ed2165d5694a414d31eac7c63ac5700fb0", "5002dd18c5940ce63b917d84f2b852c3b96009bb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 09:19:43 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 09:19:43 2010 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.open-osd.org/linux-open-osd\n\n* \u0027for-linus\u0027 of git://git.open-osd.org/linux-open-osd:\n exofs: Fix groups code when num_devices is not divisible by group_width\n exofs: Remove useless optimization\n exofs: exofs_file_fsync and exofs_file_flush correctness\n exofs: Remove superfluous dependency on buffer_head and writeback\n" }, { "commit": "4ec70c9b46b032e7f1b41b543c607d6a33b78a1a", "tree": "ab2f2509a13845130fa0a3a28f4a9944bb761e05", "parents": [ "845a2cc0507055278e0fa722ed0f8c791b7401dd" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Jun 07 11:42:26 2010 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Aug 09 16:48:24 2010 -0400" }, "message": "convert exofs to -\u003eevict_inode()\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "85dc7878c6c2277de2eda2c4d1b11ea5c5b1068a", "tree": "dde1a55f77069b824b796420b1510b2547118cd3", "parents": [ "3a09b1be53d23df780a0cd0e4087a05e2ca4a00c" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon May 31 18:55:43 2010 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Aug 04 13:17:55 2010 +0300" }, "message": "exofs: Remove superfluous dependency on buffer_head and writeback\n\nexofs_releasepage \u0026\u0026 exofs_invalidatepage are never called.\nLeave the WARN_ONs but remove any code. Remove the\n\ncleanup other stale #includes.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "b3d0ab7e60d1865bb6f6a79a77aaba22f2543236", "tree": "bd770d63b01998b8ad1ee0f9fc463465b3586e8a", "parents": [ "9df9c8b930156a2f9ce2b2ae66acb14bee2663f5" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Thu Apr 22 12:26:04 2010 +0200" }, "committer": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Thu Apr 22 12:26:04 2010 +0200" }, "message": "exofs: add bdi backing to mount session\n\nThis ensures that dirty data gets flushed properly.\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "5a0e3ad6af8660be21ca98a971cd00f331318c05", "tree": "5bfb7be11a03176a87296a43ac6647975c00a1d1", "parents": [ "ed391f4ebf8f701d3566423ce8f17e614cde9806" ], "author": { "name": "Tejun Heo", "email": "tj@kernel.org", "time": "Wed Mar 24 17:04:11 2010 +0900" }, "committer": { "name": "Tejun Heo", "email": "tj@kernel.org", "time": "Tue Mar 30 22:02:32 2010 +0900" }, "message": "include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h\n\npercpu.h is included by sched.h and module.h and thus ends up being\nincluded when building most .c files. percpu.h includes slab.h which\nin turn includes gfp.h making everything defined by the two files\nuniversally available and complicating inclusion dependencies.\n\npercpu.h -\u003e slab.h dependency is about to be removed. Prepare for\nthis change by updating users of gfp and slab facilities include those\nheaders directly instead of assuming availability. As this conversion\nneeds to touch large number of source files, the following script is\nused as the basis of conversion.\n\n http://userweb.kernel.org/~tj/misc/slabh-sweep.py\n\nThe script does the followings.\n\n* Scan files for gfp and slab usages and update includes such that\n only the necessary includes are there. ie. if only gfp is used,\n gfp.h, if slab is used, slab.h.\n\n* When the script inserts a new include, it looks at the include\n blocks and try to put the new include such that its order conforms\n to its surrounding. It\u0027s put in the include block which contains\n core kernel includes, in the same order that the rest are ordered -\n alphabetical, Christmas tree, rev-Xmas-tree or at the end if there\n doesn\u0027t seem to be any matching order.\n\n* If the script can\u0027t find a place to put a new include (mostly\n because the file doesn\u0027t have fitting include block), it prints out\n an error message indicating which .h file needs to be added to the\n file.\n\nThe conversion was done in the following steps.\n\n1. The initial automatic conversion of all .c files updated slightly\n over 4000 files, deleting around 700 includes and adding ~480 gfp.h\n and ~3000 slab.h inclusions. The script emitted errors for ~400\n files.\n\n2. Each error was manually checked. Some didn\u0027t need the inclusion,\n some needed manual addition while adding it to implementation .h or\n embedding .c file was more appropriate for others. This step added\n inclusions to around 150 files.\n\n3. The script was run again and the output was compared to the edits\n from #2 to make sure no file was left behind.\n\n4. Several build tests were done and a couple of problems were fixed.\n e.g. lib/decompress_*.c used malloc/free() wrappers around slab\n APIs requiring slab.h to be added manually.\n\n5. The script was run on all .h files but without automatically\n editing them as sprinkling gfp.h and slab.h inclusions around .h\n files could easily lead to inclusion dependency hell. Most gfp.h\n inclusion directives were ignored as stuff from gfp.h was usually\n wildly available and often used in preprocessor macros. Each\n slab.h inclusion directive was examined and added manually as\n necessary.\n\n6. percpu.h was updated not to include slab.h.\n\n7. Build test were done on the following configurations and failures\n were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my\n distributed build env didn\u0027t work with gcov compiles) and a few\n more options had to be turned off depending on archs to make things\n build (like ipr on powerpc/64 which failed due to missing writeq).\n\n * x86 and x86_64 UP and SMP allmodconfig and a custom test config.\n * powerpc and powerpc64 SMP allmodconfig\n * sparc and sparc64 SMP allmodconfig\n * ia64 SMP allmodconfig\n * s390 SMP allmodconfig\n * alpha SMP allmodconfig\n * um on x86_64 SMP allmodconfig\n\n8. percpu.h modifications were reverted so that it could be applied as\n a separate patch and serve as bisection point.\n\nGiven the fact that I had only a couple of failures from tests on step\n6, I\u0027m fairly confident about the coverage of this conversion patch.\nIf there is a breakage, it\u0027s likely to be something in one of the arch\nheaders which should be easily discoverable easily on most builds of\nthe specific arch.\n\nSigned-off-by: Tejun Heo \u003ctj@kernel.org\u003e\nGuess-its-ok-by: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nCc: Ingo Molnar \u003cmingo@redhat.com\u003e\nCc: Lee Schermerhorn \u003cLee.Schermerhorn@hp.com\u003e\n" }, { "commit": "50a76fd3c352ed2740eba01512efcfceee0703be", "tree": "425416e068648e225b41327a120d00bbddd16d0e", "parents": [ "b367e78bd1c7af4c018ce98b1f6d3e001aba895a" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Feb 11 13:01:39 2010 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Feb 28 03:55:53 2010 -0800" }, "message": "exofs: groups support\n\n* _calc_stripe_info() changes to accommodate for grouping\n calculations. Returns additional information\n\n* old _prepare_pages() becomes _prepare_one_group()\n which stores pages belonging to one device group.\n\n* New _prepare_for_striping iterates on all groups calling\n _prepare_one_group().\n\n* Enable mounting of groups data_maps (group_width !\u003d 0)\n\n[QUESTION]\nwhat is faster A or B;\nA.\tx +\u003d stride;\n\tx \u003d x % width + first_x;\n\nB\tx +\u003d stride\n\tif (x \u003c last_x)\n\t\tx \u003d first_x;\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "5d952b8391692553c31e620a92d6e09262a9a307", "tree": "b3a1a0490fc98b6304685d64bb4774235ec94a2d", "parents": [ "d9c740d2253e75db8cef8f87a3125c450f3ebd82" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Feb 01 13:35:51 2010 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Feb 28 03:43:08 2010 -0800" }, "message": "exofs: RAID0 support\n\nWe now support striping over mirror devices. Including variable sized\nstripe_unit.\n\nSome limits:\n* stripe_unit must be a multiple of PAGE_SIZE\n* stripe_unit * stripe_count is maximum upto 32-bit (4Gb)\n\nTested RAID0 over mirrors, RAID0 only, mirrors only. All check.\n\nDesign notes:\n* I\u0027m not using a vectored raid-engine mechanism yet. Following the\n pnfs-objects-layout data-map structure, \"Mirror\" is just a private\n case of \"group_width\" \u003d\u003d 1, and RAID0 is a private case of\n \"Mirrors\" \u003d\u003d 1. The performance lose of the general case over the\n particular special case optimization is totally negligible, also\n considering the extra code size.\n\n* In general I added a prepare_stripes() stage that divides the\n to-be-io pages to the participating devices, the previous\n exofs_ios_write/read, now becomes _write/read_mirrors and a new\n write/read upper layer loops on all devices calling\n _write/read_mirrors. Effectively the prepare_stripes stage is the all\n secret.\n Also truncate need fixing to accommodate for striping.\n\n* In a RAID0 arrangement, in a regular usage scenario, if all inode\n layouts will start at the same device, the small files fill up the\n first device and the later devices stay empty, the farther the device\n the emptier it is.\n\n To fix that, each inode will start at a different stripe_unit,\n according to it\u0027s obj_id modulus number-of-stripe-units. And\n will then span all stripe-units in the same incrementing order\n wrapping back to the beginning of the device table. We call it\n a stripe-units moving window.\n\n Special consideration was taken to keep all devices in a mirror\n arrangement identical. So a broken osd-device could just be cloned\n from one of the mirrors and no FS scrubbing is needed. (We do that\n by rotating stripe-unit at a time and not a single device at a time.)\n\nTODO:\n We no longer verify object_length \u003d\u003d inode-\u003ei_size in exofs_iget.\n (since i_size is stripped on multiple objects now).\n I should introduce a multiple-device attribute reading, and use\n it in exofs_iget.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "45d3abcb1a7388b2b97582e13bf9dd21784dcaa5", "tree": "e957fdec303c073490c261999d6939c2ecf20352", "parents": [ "22ddc556380cf5645c52292b6d980766646eb864" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Jan 28 11:46:16 2010 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Feb 28 03:35:27 2010 -0800" }, "message": "exofs: Move layout related members to a layout structure\n\n* Abstract away those members in exofs_sb_info that are related/needed\n by a layout into a new exofs_layout structure. Embed it in exofs_sb_info.\n\n* At exofs_io_state receive/keep a pointer to an exofs_layout. No need for\n an exofs_sb_info pointer, all we need is at exofs_layout.\n\n* Change any usage of above exofs_sb_info members to their new name.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "04dc1e88ad9c9f9639019e9646a89ce0ebf706bb", "tree": "403206d1e85e9e487d847694cbe0ecf111b3f02b", "parents": [ "06886a5a3dc5a5abe0a4d257c26317bde7047be8" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Nov 16 16:03:05 2009 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Dec 10 09:59:23 2009 +0200" }, "message": "exofs: Multi-device mirror support\n\nThis patch changes on-disk format, it is accompanied with a parallel\npatch to mkfs.exofs that enables multi-device capabilities.\n\nAfter this patch, old exofs will refuse to mount a new formatted FS and\nnew exofs will refuse an old format. This is done by moving the magic\nfield offset inside the FSCB. A new FSCB *version* field was added. In\nthe future, exofs will refuse to mount unmatched FSCB version. To\nup-grade or down-grade an exofs one must use mkfs.exofs --upgrade option\nbefore mounting.\n\nIntroduced, a new object that contains a *device-table*. This object\ncontains the default *data-map* and a linear array of devices\ninformation, which identifies the devices used in the filesystem. This\nobject is only written to offline by mkfs.exofs. This is why it is kept\nseparate from the FSCB, since the later is written to while mounted.\n\nSame partition number, same object number is used on all devices only\nthe device varies.\n\n* define the new format, then load the device table on mount time make\n sure every thing is supported.\n\n* Change I/O engine to now support Mirror IO, .i.e write same data\n to multiple devices, read from a random device to spread the\n read-load from multiple clients (TODO: stripe read)\n\nImplementation notes:\n A few points introduced in previous patch should be mentioned here:\n\n* Special care was made so absolutlly all operation that have any chance\n of failing are done before any osd-request is executed. This is to\n minimize the need for a data consistency recovery, to only real IO\n errors.\n\n* Each IO state has a kref. It starts at 1, any osd-request executed\n will increment the kref, finally when all are executed the first ref\n is dropped. At IO-done, each request completion decrements the kref,\n the last one to return executes the internal _last_io() routine.\n _last_io() will call the registered io_state_done. On sync mode a\n caller does not supply a done method, indicating a synchronous\n request, the caller is put to sleep and a special io_state_done is\n registered that will awaken the caller. Though also in sync mode all\n operations are executed in parallel.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "06886a5a3dc5a5abe0a4d257c26317bde7047be8", "tree": "858ac56e120c0473d764fc64a2660e6d79729c8c", "parents": [ "8ce9bdd1fbe962933736d7977e972972cd5d754c" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Nov 08 14:54:08 2009 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Dec 10 09:59:22 2009 +0200" }, "message": "exofs: Move all operations to an io_engine\n\nIn anticipation for multi-device operations, we separate osd operations\ninto an abstract I/O API. Currently only one device is used but later\nwhen adding more devices, we will drive all devices in parallel according\nto a \"data_map\" that describes how data is arranged on multiple devices.\nThe file system level operates, like before, as if there is one object\n(inode-number) and an i_size. The io engine will split this to the same\nobject-number but on multiple device.\n\nAt first we introduce Mirror (raid 1) layout. But at the final outcome\nwe intend to fully implement the pNFS-Objects data-map, including\nraid 0,4,5,6 over mirrored devices, over multiple device-groups. And\nmore. See: http://tools.ietf.org/html/draft-ietf-nfsv4-pnfs-obj-12\n\n* Define an io_state based API for accessing osd storage devices\n in an abstract way.\n Usage:\n\tFirst a caller allocates an io state with:\n\t\texofs_get_io_state(struct exofs_sb_info *sbi,\n\t\t\t\t struct exofs_io_state** ios);\n\n\tThen calles one of:\n\t\texofs_sbi_create(struct exofs_io_state *ios);\n\t\texofs_sbi_remove(struct exofs_io_state *ios);\n\t\texofs_sbi_write(struct exofs_io_state *ios);\n\t\texofs_sbi_read(struct exofs_io_state *ios);\n\t\texofs_oi_truncate(struct exofs_i_info *oi, u64 new_len);\n\n\tAnd when done\n\t\texofs_put_io_state(struct exofs_io_state *ios);\n\n* Convert all source files to use this new API\n* Convert from bio_alloc to bio_kmalloc\n* In io engine we make use of the now fixed osd_req_decode_sense\n\nThere are no functional changes or on disk additions after this patch.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "cae012d8532879544326fff5fa2ae22a6dfe8e23", "tree": "7958c2c11cee1b0c342f7902cd538dcec679989e", "parents": [ "19fe294f2eaee33574ac1fdcf3cc26880de820ea" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Mon Nov 02 18:19:24 2009 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Dec 10 09:59:21 2009 +0200" }, "message": "exofs: statfs blocks is sectors not FS blocks\n\nEven though exofs has a 4k block size, statfs blocks\nis in sectors (512 bytes).\n\nAlso if target returns 0 for capacity then make it\nULLONG_MAX. df does not like zero-size filesystems\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "19fe294f2eaee33574ac1fdcf3cc26880de820ea", "tree": "ea00168d7f7fb8594c40a41cdb9d4aeea8ec9222", "parents": [ "9cfdc7aa9f1b59627029ad00a58c3f59eb2cc383" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Sep 03 20:38:02 2009 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Thu Dec 10 09:59:20 2009 +0200" }, "message": "exofs: Prints on mount and unmout\n\nIt is important to print in the logs when a filesystem was\nmounted and eventually unmounted.\n\nPrint the osd-device\u0027s osd_name and pid the FS was\nmounted/unmounted on.\n\nTODO: How to also print the namespace path the filesystem was\n mounted on?\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "1ba50bbe93ebb98e83b174a85eff76af430c4e5b", "tree": "4c7b7db057fa5ca60a4e1adcb5dfb63cbf5a5702", "parents": [ "88a0a53d702b1fa39ed9e631939d2dbd92dfe486" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Wed Aug 19 17:56:46 2009 +0300" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Thu Sep 24 07:47:38 2009 -0400" }, "message": "exofs: remove BKL from super operations\n\nthe two places inside exofs that where taking the BKL were:\nexofs_put_super() - .put_super\nand\nexofs_sync_fs() - which is .sync_fs and is also called from\n .write_super.\n\nNow exofs_sync_fs() is protected from itself by also taking\nthe sb_lock.\n\nexofs_put_super() directly calls exofs_sync_fs() so there is no\ndanger between these two either.\n\nIn anyway there is absolutely nothing dangerous been done\ninside exofs_sync_fs().\n\nUnless there is some subtle race with the actual lifetime of\nthe super_block in regard to .put_super and some other parts\nof the VFS. Which is highly unlikely.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "405f55712dfe464b3240d7816cc4fe4174831be2", "tree": "96c425ea7fa8b31058b8f83a433c5e5265c8ebc7", "parents": [ "f9fabcb58a6d26d6efde842d1703ac7cfa9427b6" ], "author": { "name": "Alexey Dobriyan", "email": "adobriyan@gmail.com", "time": "Sat Jul 11 22:08:37 2009 +0400" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 12 12:22:34 2009 -0700" }, "message": "headers: smp_lock.h redux\n\n* Remove smp_lock.h from files which don\u0027t need it (including some headers!)\n* Add smp_lock.h to files which do need it\n* Make smp_lock.h include conditional in hardirq.h\n It\u0027s needed only for one kernel_locked() usage which is under CONFIG_PREEMPT\n\n This will make hardirq.h inclusion cheaper for every PREEMPT\u003dn config\n (which includes allmodconfig/allyesconfig, BTW)\n\nSigned-off-by: Alexey Dobriyan \u003cadobriyan@gmail.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "baaf94cdc7fe1c61e3c660a3b055724fd9d0a034", "tree": "b6f1704f38e39801df39b231e19b514d3b4b0ddf", "parents": [ "27d2e1491985e95c486d991302e399f5c584b4eb" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Jun 14 16:52:10 2009 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Jun 21 17:53:47 2009 +0300" }, "message": "exofs: Avoid using file_fsync()\n\nThe use of file_fsync() in exofs_file_sync() is not necessary since it\ndoes some extra stuff not used by exofs. Open code just the parts that\nare currently needed.\n\nTODO: Farther optimization can be done to sync the sb only on inode\nupdate of new files, Usually the sb update is not needed in exofs.\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "27d2e1491985e95c486d991302e399f5c584b4eb", "tree": "7f978b04cbabebe44108f246d7002c60ad9fdd35", "parents": [ "b76a3f93d01fc93a87cb6eba4e854ffe378b4bac" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Jun 14 17:23:09 2009 +0300" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Jun 21 17:53:47 2009 +0300" }, "message": "exofs: Remove IBM copyrights\n\nBoaz,\nCongrats on getting all the OSD stuff into 2.6.30!\nI just pulled the git, and saw that the IBM copyrights are still there.\nPlease remove them from all files:\n * Copyright (C) 2005, 2006\n * International Business Machines\n\nIBM has revoked all rights on the code - they gave it to me.\n\nThanks!\nAvishay\n\nSigned-off-by: Avishay Traeger \u003cavishay@gmail.com\u003e\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "80e09fb942d38beb19dcffbeb14d496beeb0a989", "tree": "cdaa5cf4baf745ee4c83b8e03cb75581e7ea1ffa", "parents": [ "561e47ce7244168788d4ecef9a2271df204b3c89" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Mon Jun 08 10:03:58 2009 +0200" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Thu Jun 11 21:36:15 2009 -0400" }, "message": "exofs: add -\u003esync_fs\n\nAdd a -\u003esync_fs method for data integrity syncs, and reimplement\n-\u003ewrite_super ontop of it.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "ebc1ac164560a241d9bf1b7519062910c3f90a01", "tree": "db17fd3043b15f8269cce2c2d00213814022fc30", "parents": [ "01ba687577647beef6c5f2ea59bfb56fac9fcde2" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Mon May 11 23:35:03 2009 +0200" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Thu Jun 11 21:36:09 2009 -0400" }, "message": "-\u003ewrite_super lock_super pushdown\n\nPush down lock_super into -\u003ewrite_super instances and remove it from the\ncaller.\n\nFollowing filesystem don\u0027t need -\u003es_lock in -\u003ewrite_super and are skipped:\n\n * bfs, nilfs2 - no other uses of s_lock and have internal locks in\n\t-\u003ewrite_super\n * ext2 - uses BKL in ext2_write_super and has internal calls without s_lock\n * reiserfs - no other uses of s_lock as has reiserfs_write_lock (BKL) in\n \t-\u003ewrite_super\n * xfs - no other uses of s_lock and uses internal lock (buffer lock on\n\tsuperblock buffer) to serialize -\u003ewrite_super. Also xfs_fs_write_super\n\tis superflous and will go away in the next merge window\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "6cfd0148425e528b859b26e436b01f23f6926224", "tree": "60e3257053554ff198fe5825e6f12a00c3b4422a", "parents": [ "a9e220f8322e2b0e0b8903fe00265461cffad3f0" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Tue May 05 15:40:36 2009 +0200" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Thu Jun 11 21:36:07 2009 -0400" }, "message": "push BKL down into -\u003eput_super\n\nMove BKL into -\u003eput_super from the only caller. A couple of\nfilesystems had trivial enough -\u003eput_super (only kfree and NULLing of\ns_fs_info + stuff in there) to not get any locking: coda, cramfs, efs,\nhugetlbfs, omfs, qnx4, shmem, all others got the full treatment. Most\nof them probably don\u0027t need it, but I\u0027d rather sort that out individually.\nPreferably after all the other BKL pushdowns in that area.\n\n[AV: original used to move lock_super() down as well; these changes are\nremoved since we don\u0027t do lock_super() at all in generic_shutdown_super()\nnow]\n[AV: fuse, btrfs and xfs are known to need no damn BKL, exempt]\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "8c85e125124a473d6f3e9bb187b0b84207f81d91", "tree": "d883424a156c00c55193e46567efdffddb6ef6b9", "parents": [ "517bfae28353e996160518add4d00033d3886e61" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Tue Apr 28 18:00:26 2009 +0200" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Thu Jun 11 21:36:06 2009 -0400" }, "message": "remove -\u003ewrite_super call in generic_shutdown_super\n\nWe just did a full fs writeout using sync_filesystem before, and if\nthat\u0027s not enough for the filesystem it can perform it\u0027s own writeout\nin -\u003eput_super, which many filesystems already do.\n\nMove a call to foofs_write_super into every foofs_put_super for now to\nguarantee identical behaviour until it\u0027s cleaned up by the individual\nfilesystem maintainers.\n\nExceptions:\n\n - affs already has identical copy \u0026 pasted code at the beginning of\n affs_put_super so no need to do it twice.\n - xfs does the right thing without it and I have changes pending for\n the xfs tree touching this are so I don\u0027t really need conflicts\n here..\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "8cf74b3936f5c65975b68333727f0540ec8350f4", "tree": "99eda92b32dc7869d6f3a28215f5b854675db6b1", "parents": [ "ba9e5e98ca2f808fe92b103a8e6ce5271b10cc89" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Sun Mar 22 12:47:26 2009 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 31 19:44:36 2009 +0300" }, "message": "exofs: export_operations\n\nimplement export_operations and set in superblock.\nIt is now posible to export exofs via nfs\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" }, { "commit": "ba9e5e98ca2f808fe92b103a8e6ce5271b10cc89", "tree": "0ec6eb23adfe0f762585e3a0c0f559bd91b74b11", "parents": [ "e6af00f1d1697ca41ab6a55307066ef3466833a9" ], "author": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Oct 28 16:11:41 2008 +0200" }, "committer": { "name": "Boaz Harrosh", "email": "bharrosh@panasas.com", "time": "Tue Mar 31 19:44:34 2009 +0300" }, "message": "exofs: super_operations and file_system_type\n\nThis patch ties all operation vectors into a file system superblock\nand registers the exofs file_system_type at module\u0027s load time.\n\n* The file system control block (AKA on-disk superblock) resides in\n an object with a special ID (defined in common.h).\n Information included in the file system control block is used to\n fill the in-memory superblock structure at mount time. This object\n is created before the file system is used by mkexofs.c It contains\n information such as:\n\t- The file system\u0027s magic number\n\t- The next inode number to be allocated\n\nSigned-off-by: Boaz Harrosh \u003cbharrosh@panasas.com\u003e\n" } ] }