)]}' { "log": [ { "commit": "be867b194a3ae3c680c29521287ae49b4d44d420", "tree": "1b3ad125d02603c6a9cc4bb8ae518695d7c6a66d", "parents": [ "f349cf34731c3b3eddbcd63cb04489aee2c405bd" ], "author": { "name": "Serge E. Hallyn", "email": "serue@us.ibm.com", "time": "Mon May 03 16:23:15 2010 -0500" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri May 21 09:37:31 2010 -0700" }, "message": "sysfs: Comment sysfs directory tagging logic\n\nAdd some in-line comments to explain the new infrastructure, which\nwas introduced to support sysfs directory tagging with namespaces.\nI think an overall description someplace might be good too, but it\ndidn\u0027t really seem to fit into Documentation/filesystems/sysfs.txt,\nwhich appears more geared toward users, rather than maintainers, of\nsysfs.\n\n(Tejun, please let me know if I can make anything clearer or failed\naltogether to comment something that should be commented.)\n\nSigned-off-by: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nCc: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "3ff195b011d7decf501a4d55aeed312731094796", "tree": "8cfdc330abbf82893955f2d7d6e96efee81bfd7c", "parents": [ "bc451f2058238013e1cdf4acd443c01734d332f0" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Tue Mar 30 11:31:26 2010 -0700" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri May 21 09:37:31 2010 -0700" }, "message": "sysfs: Implement sysfs tagged directory support.\n\nThe problem. When implementing a network namespace I need to be able\nto have multiple network devices with the same name. Currently this\nis a problem for /sys/class/net/*, /sys/devices/virtual/net/*, and\npotentially a few other directories of the form /sys/ ... /net/*.\n\nWhat this patch does is to add an additional tag field to the\nsysfs dirent structure. For directories that should show different\ncontents depending on the context such as /sys/class/net/, and\n/sys/devices/virtual/net/ this tag field is used to specify the\ncontext in which those directories should be visible. Effectively\nthis is the same as creating multiple distinct directories with\nthe same name but internally to sysfs the result is nicer.\n\nI am calling the concept of a single directory that looks like multiple\ndirectories all at the same path in the filesystem tagged directories.\n\nFor the networking namespace the set of directories whose contents I need\nto filter with tags can depend on the presence or absence of hotplug\nhardware or which modules are currently loaded. Which means I need\na simple race free way to setup those directories as tagged.\n\nTo achieve a reace free design all tagged directories are created\nand managed by sysfs itself.\n\nUsers of this interface:\n- define a type in the sysfs_tag_type enumeration.\n- call sysfs_register_ns_types with the type and it\u0027s operations\n- sysfs_exit_ns when an individual tag is no longer valid\n\n- Implement mount_ns() which returns the ns of the calling process\n so we can attach it to a sysfs superblock.\n- Implement ktype.namespace() which returns the ns of a syfs kobject.\n\nEverything else is left up to sysfs and the driver layer.\n\nFor the network namespace mount_ns and namespace() are essentially\none line functions, and look to remain that.\n\nTags are currently represented a const void * pointers as that is\nboth generic, prevides enough information for equality comparisons,\nand is trivial to create for current users, as it is just the\nexisting namespace pointer.\n\nThe work needed in sysfs is more extensive. At each directory\nor symlink creating I need to check if the directory it is being\ncreated in is a tagged directory and if so generate the appropriate\ntag to place on the sysfs_dirent. Likewise at each symlink or\ndirectory removal I need to check if the sysfs directory it is\nbeing removed from is a tagged directory and if so figure out\nwhich tag goes along with the name I am deleting.\n\nCurrently only directories which hold kobjects, and\nsymlinks are supported. There is not enough information\nin the current file attribute interfaces to give us anything\nto discriminate on which makes it useless, and there are\nno potential users which makes it an uninteresting problem\nto solve.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Benjamin Thery \u003cbenjamin.thery@bull.net\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "9e7fdd25b21ebf50713dfb7702af57c270e5572a", "tree": "ea1c4d9c33be9b72782081fc16a5bac8aafe93da", "parents": [ "c3473c63542d53740f175f3a515257ae159e998b" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Tue Mar 30 11:31:24 2010 -0700" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri May 21 09:37:30 2010 -0700" }, "message": "sysfs: Basic support for multiple super blocks\n\nAdd all of the necessary bioler plate to support\nmultiple superblocks in sysfs.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "0f4288ec6fcc1a47d1fa0241ec1c6dacd5a09e96", "tree": "7112809d470a7e6dc55cd68054896beef24bc184", "parents": [ "fac2622bbad4d7e3a32c53e631e018b80ec631dc" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Feb 12 19:22:28 2010 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Sun Mar 07 17:04:52 2010 -0800" }, "message": "sysfs: Kill unused sysfs_sb variable.\n\nNow that there are no more users we can remove\nthe sysfs_sb variable.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "fac2622bbad4d7e3a32c53e631e018b80ec631dc", "tree": "42e4ad2b38c6d40c86eabaff5efff2b0adeb0b40", "parents": [ "2354dcc7218853a6537ec722be40fde9a11c413b" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Feb 12 19:22:27 2010 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Sun Mar 07 17:04:52 2010 -0800" }, "message": "sysfs: Pass super_block to sysfs_get_inode\n\nCurrently sysfs_get_inode magically returns an inode on\nsysfs_sb. Make the super_block parameter explicit and\nthe code becomes clearer.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "19c38b632d4554329986021fd66bdea468cf452c", "tree": "a724445bc46a4500ff6c4def2a6de39fc438f8b5", "parents": [ "f8d4f618fed5a4978afada52166bc2efaf7656d1" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@aristanetworks.com", "time": "Fri Feb 12 19:22:24 2010 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Sun Mar 07 17:04:52 2010 -0800" }, "message": "sysfs: Pack sysfs_dirent more tightly.\n\nPlacing the 16bit s_mode between a pointer and a long doesn\u0027t pack well\nespecailly on 64bit where we wast 48 bits. So move s_mode and\ndeclare it as a unsigned short. This is the sysfs backing store\nafter all we don\u0027t need fields extra large just in case someday\nwe want userspace to be able to use a larger value.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "6992f5334995af474c2b58d010d08bc597f0f2fe", "tree": "bef650bfb9355b8d7cbb7a5acb3632c02ec7fbe7", "parents": [ "a2db6842873c8e5a70652f278d469128cb52db70" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Thu Feb 11 15:21:53 2010 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Sun Mar 07 17:04:51 2010 -0800" }, "message": "sysfs: Use one lockdep class per sysfs attribute.\n\nAcknowledge that the logical sysfs rwsem has one instance per\nsysfs attribute with different locking depencencies for different\nattributes.\n\nThere is a sysfs idiom where writing to one sysfs file causes the\naddition or removal of other sysfs files. Lumping all of the\nsysfs attributes together in one lock class causes lockdep to\ngenerate lots of false positives.\n\nThis introduces the requirement that non-static sysfs attributes\nneed to be initialized with sysfs_attr_init or sysfs_bin_attr_init.\nStrictly speaking this requirement only exists when lockdep is\nenabled, and when lockdep is enabled we get a bit fat warning\nif this requirement is not met.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nAcked-by: WANG Cong \u003cxiyou.wangcong@gmail.com\u003e\nCc: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "a2db6842873c8e5a70652f278d469128cb52db70", "tree": "0415c6cfb33753c7a20e05d95e3ce485a60b1f41", "parents": [ "e72ceb8ccac5f770b3e696e09bb673dca7024b20" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Thu Feb 11 15:20:00 2010 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Sun Mar 07 17:04:51 2010 -0800" }, "message": "sysfs: Only take active references on attributes.\n\nIf we exclude directories and symlinks from the set of sysfs\ndirents where we need active references we are left with\nsysfs attributes (binary or not).\n\n- Tweak sysfs_deactivate to only do something on attributes\n- Move lockdep initialization into sysfs_file_add_mode to\n limit it to just attributes.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nAcked-by: WANG Cong \u003cxiyou.wangcong@gmail.com\u003e\nCc: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "e72ceb8ccac5f770b3e696e09bb673dca7024b20", "tree": "9868803df687838c3c5f6f2265ceb7532b93a5f4", "parents": [ "3c31f07ad0dab02fe17195d32a965d57fd947707" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Thu Feb 11 15:18:38 2010 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Sun Mar 07 17:04:51 2010 -0800" }, "message": "sysfs: Remove sysfs_get/put_active_two\n\nIt turns out that holding an active reference on a directory is\npointless. The purpose of the active references are to allows us to\nblock when removing sysfs entries that have custom methods so we don\u0027t\nremove modules while running modular code and to keep those custom\nmethods from accessing data structures after the files have been\nremoved. Further sysfs_remove_dir remove all elements in the\ndirectory before removing the directory itself, so there is no chance\nwe will remove a directory with active children.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nCc: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "846f99749ab68bbc7f75c74fec305de675b1a1bf", "tree": "4c1ca68cfea804bd5215d57844a16f69c566aedc", "parents": [ "3e27249c84beed1c79d767b350e52ad038db9053" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Sat Jan 02 13:37:12 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jan 04 12:34:46 2010 -0800" }, "message": "sysfs: Add lockdep annotations for the sysfs active reference\n\nHolding locks over device_del -\u003e kobject_del -\u003e sysfs_deactivate can\ncause deadlocks if those same locks are grabbed in sysfs show or store\nmethods.\n\nThe I model s_active count + completion as a sleeping read/write lock.\nI describe to lockdep sysfs_get_active as a read_trylock,\nsysfs_put_active as a read_unlock, and sysfs_deactivate as a\nwrite_lock and write_unlock pair. This seems to capture the essence\nfor purposes of finding deadlocks, and in my testing gives finds real\nissues and ignores non-issues.\n\nThis brings us back to holding locks over kobject_del is a problem\nthat ideally we should find a way of addressing, but at least lockdep\ncan tell us about the problems instead of requiring developers to debug\nrare strange system deadlocks, that happen when sysfs files are removed\nwhile being written to.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ca1bab38195d66bdf42320a99cc7359434a271d3", "tree": "4394403b4a85fb269d052b7852b41ee0f7ba8a88", "parents": [ "832b6af198aefe6034310e124594cc8b833c0ef9" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Nov 20 16:08:57 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Dec 11 11:24:54 2009 -0800" }, "message": "sysfs: Factor out sysfs_rename from sysfs_rename_dir and sysfs_move_dir\n\nThese two functions do 90% of the same work and it doesn\u0027t significantly\nobfuscate the function to allow both the parent dir and the name to change\nat the same time. So merge them together to simplify maintenance, and\nincrease testing.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "832b6af198aefe6034310e124594cc8b833c0ef9", "tree": "853eec8f7449dd7fc51c6d4c9b70ea44b81cf0d9", "parents": [ "a16bbc3430ed94b543222f4c8ef68025f8493e93" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Nov 20 16:08:56 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Dec 11 11:24:54 2009 -0800" }, "message": "sysfs: Propagate renames to the vfs on demand\n\nBy teaching sysfs_revalidate to hide a dentry for\na sysfs_dirent if the sysfs_dirent has been renamed,\nand by teaching sysfs_lookup to return the original\ndentry if the sysfs dirent has been renamed. I can\nshow the results of renames correctly without having to\nupdate the dcache during the directory rename.\n\nThis massively simplifies the rename logic allowing a lot\nof weird sysfs special cases to be removed along with\na lot of now unnecesary helper code.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "a16bbc3430ed94b543222f4c8ef68025f8493e93", "tree": "e866babd9a5ab024bcd68a9f935c4b373b1e8ac7", "parents": [ "06fc0d66f7ed3a3b08e8fcf8c325ecf0b8f93fea" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@maxwell.aristanetworks.com", "time": "Fri Nov 20 16:08:55 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Dec 11 11:24:54 2009 -0800" }, "message": "sysfs: Gut sysfs_addrm_start and sysfs_addrm_finish\n\nWith lazy inode updates and dentry operations bringing everything\ninto sync on demand there is no longer any need to immediately\nupdate the vfs or grab i_mutex to protect those updates as we\nmake changes to sysfs.\n\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "e61ab4ae48fbf477f5b9fcbec9e1b8dc789920d0", "tree": "70199a29a97a4f5ba8aea4340bc828da4f550e5d", "parents": [ "c099aacd48ee73bd2de7da029e536ed005d72a43" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Nov 20 16:08:53 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Dec 11 11:24:54 2009 -0800" }, "message": "sysfs: Implement sysfs_getattr \u0026 sysfs_permission\n\nWith the implementation of sysfs_getattr and sysfs_permission\nsysfs becomes able to lazily propogate inode attribute changes\nfrom the sysfs_dirents to the vfs inodes. This paves the way\nfor deleting significant chunks of now unnecessary code.\n\nWhile doing this we did not reference sysfs_setattr from\nsysfs_symlink_inode_operations so I added along with\nsysfs_getattr and sysfs_permission.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "35df63c46c8605d00140ec4833d2104e2e9f1acc", "tree": "1f02d75ab4589d3c8878e38e9d45360a0c0667d4", "parents": [ "4be3df28beec5605c77a18aa2a4f987b5648f9ce" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Fri Nov 20 16:08:50 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Dec 11 11:24:53 2009 -0800" }, "message": "sysfs: Fix locking and factor out sysfs_sd_setattr\n\nCleanly separate the work that is specific to setting the\nattributes of a sysfs_dirent from what is needed to update\nthe attributes of a vfs inode.\n\nAdditionally grab the sysfs_mutex to keep any nasties from\nsurprising us when updating the sysfs_dirent.\n\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "ddd29ec6597125c830f7badb608a86c98b936b64", "tree": "e6df1ef9a635179de78650d006ecb4cd1453ebb1", "parents": [ "1ee65e37e904b959c24404139f5752edc66319d5" ], "author": { "name": "David P. Quigley", "email": "dpquigl@tycho.nsa.gov", "time": "Wed Sep 09 14:25:37 2009 -0400" }, "committer": { "name": "James Morris", "email": "jmorris@namei.org", "time": "Thu Sep 10 10:11:29 2009 +1000" }, "message": "sysfs: Add labeling support for sysfs\n\nThis patch adds a setxattr handler to the file, directory, and symlink\ninode_operations structures for sysfs. The patch uses hooks introduced in the\nprevious patch to handle the getting and setting of security information for\nthe sysfs inodes. As was suggested by Eric Biederman the struct iattr in the\nsysfs_dirent structure has been replaced by a structure which contains the\niattr, secdata and secdata length to allow the changes to persist in the event\nthat the inode representing the sysfs_dirent is evicted. Because sysfs only\nstores this information when a change is made all the optional data is moved\ninto one dynamically allocated field.\n\nThis patch addresses an issue where SELinux was denying virtd access to the PCI\nconfiguration entries in sysfs. The lack of setxattr handlers for sysfs\nrequired that a single label be assigned to all entries in sysfs. Granting virtd\naccess to every entry in sysfs is not an acceptable solution so fine grained\nlabeling of sysfs is required such that individual entries can be labeled\nappropriately.\n\n[sds: Fixed compile-time warnings, coding style, and setting of inode security init flags.]\n\nSigned-off-by: David P. Quigley \u003cdpquigl@tycho.nsa.gov\u003e\nSigned-off-by: Stephen D. Smalley \u003csds@tycho.nsa.gov\u003e\nSigned-off-by: James Morris \u003cjmorris@namei.org\u003e\n" }, { "commit": "e0edd3c65aa5b53e20280565a7ce11675eb7ed6b", "tree": "e9d7a8834eba3f19635f770f5d19bfe88a2485e4", "parents": [ "ae1b41715ee2aae356fbcca032838b71d70b855f" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@aristanetworks.com", "time": "Wed Mar 04 11:57:20 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Tue Mar 24 16:38:26 2009 -0700" }, "message": "sysfs: don\u0027t block indefinitely for unmapped files.\n\nModify sysfs bin files so that we can remove the bin file while they are\nstill mapped. When the kobject is removed we unmap the bin file and\narrange for future accesses to the mapping to receive SIGBUS.\n\nImplementing this prevents a nasty DOS when pci devices are hot plugged\nand unplugged. Where if any of their resources were mmaped the kernel\ncould not free up their pci resources or release their pci data\nstructures.\n\n[akpm@linux-foundation.org: remove unused var]\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nCc: Jesse Barnes \u003cjbarnes@virtuousgeek.org\u003e\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nCc: Kay Sievers \u003ckay.sievers@vrfy.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "04256b4a8fc73f54cd14f20867882c299728a446", "tree": "16267f6c77e0fa9f7229a30f66343289b6d1108a", "parents": [ "425cb02912d1095febfeaf8d379af7b2ac9e4a89" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@aristanetworks.com", "time": "Wed Feb 11 13:20:23 2009 -0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Tue Mar 24 16:38:25 2009 -0700" }, "message": "sysfs: reference sysfs_dirent from sysfs inodes\n\nThe sysfs_dirent serves as both an inode and a directory entry\nfor sysfs. To prevent the sysfs inode numbers from being freed\nprematurely hold a reference to sysfs_dirent from the sysfs inode.\n\n[akpm@linux-foundation.org: add comment]\nSigned-off-by: Eric W. Biederman \u003cebiederm@aristanetworks.com\u003e\nCc: Tejun Heo \u003ctj@kernel.org\u003e\nCc: Al Viro \u003cviro@ZenIV.linux.org.uk\u003e\nCc: Cornelia Huck \u003ccornelia.huck@de.ibm.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": "f1282c844e86db5a041afa41335b5f9eea6cec0c", "tree": "3736285f2f7ce145fb06538d616a9c1165ffc125", "parents": [ "ec748fa9ed3fec44aeebbf86ae050b0cc7a978d9" ], "author": { "name": "Neil Brown", "email": "neilb@suse.de", "time": "Wed Jul 16 08:58:04 2008 +1000" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Thu Oct 16 09:24:47 2008 -0700" }, "message": "sysfs: Support sysfs_notify from atomic context with new sysfs_notify_dirent\n\nSupport sysfs_notify from atomic context with new sysfs_notify_dirent\n\nsysfs_notify currently takes sysfs_mutex.\nThis means that it cannot be called in atomic context.\nsysfs_mutex is sometimes held over a malloc (sysfs_rename_dir)\nso it can block on low memory.\n\nIn md I want to be able to notify on a sysfs attribute from\natomic context, and I don\u0027t want to block on low memory because I\ncould be in the writeout path for freeing memory.\n\nSo:\n - export the \"sysfs_dirent\" structure along with sysfs_get, sysfs_put\n and sysfs_get_dirent so I can get the sysfs_dirent that I want to\n notify on and hold it in an md structure.\n - split sysfs_notify_dirent out of sysfs_notify so the sysfs_dirent\n can be notified on with no blocking (just a spinlock).\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nAcked-by: Tejun Heo \u003ctj@kernel.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "36ce6dad6e3cb3f050ed41e0beac0070d2062b25", "tree": "91c89c903b06dc8b76e66e7b2341bcd6085e81d0", "parents": [ "0ad1d6f37cc3bb234c6e7ae30e40d1d40b9aa258" ], "author": { "name": "Cornelia Huck", "email": "cornelia.huck@de.ibm.com", "time": "Tue Jun 10 11:09:08 2008 +0200" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Mon Jul 21 21:55:01 2008 -0700" }, "message": "driver core: Suppress sysfs warnings for device_rename().\n\ndriver core: Suppress sysfs warnings for device_rename().\n\nRenaming network devices to an already existing name is not\nsomething we want sysfs to print a scary warning for, since the\ncallers can deal with this correctly. So let\u0027s introduce\nsysfs_create_link_nowarn() which gets rid of the common warning.\n\nSigned-off-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "0f4238958d28044b335644b69df6071cdb04b5ce", "tree": "159b82f8eb9479cdc10f861d682309b4c0c8c411", "parents": [ "bbd1ae412c9eb09ae7bb11cfaf7018a2367d493f" ], "author": { "name": "James Bottomley", "email": "James.Bottomley@HansenPartnership.com", "time": "Thu Mar 20 20:47:52 2008 -0500" }, "committer": { "name": "James Bottomley", "email": "James.Bottomley@HansenPartnership.com", "time": "Tue Apr 22 15:16:31 2008 -0500" }, "message": "[SCSI] sysfs: make group is_valid return a mode_t\n\nWe have a problem in scsi_transport_spi in that we need to customise\nnot only the visibility of the attributes, but also their mode. Fix\nthis by making the is_visible() callback return a mode, with 0\nindicating is not visible.\n\nAlso add a sysfs_update_group() API to allow us to change either the\nvisibility or mode of the files at any time on the fly.\n\nAcked-by: Kay Sievers \u003ckay.sievers@vrfy.org\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@HansenPartnership.com\u003e\n" }, { "commit": "78e9d3678c8362aad2b2a48c242966aebb089dbd", "tree": "8521c590a8682fb6039e66990213c292068b9475", "parents": [ "34358c26a2c96b2a068dc44e0ac602106a466bce" ], "author": { "name": "Adrian Bunk", "email": "bunk@kernel.org", "time": "Wed Oct 24 18:23:32 2007 +0200" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Tue Oct 30 21:52:33 2007 -0700" }, "message": "sysfs: make sysfs_{get,put}_active() static\n\nsysfs_{get,put}_active() can now become static.\n\nSigned-off-by: Adrian Bunk \u003cbunk@kernel.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "e0bf68ddec4f4f90e5871404be4f1854c17f3120", "tree": "36203a3558cbe26d698bed18be69b3822fb5eef2", "parents": [ "dc62a30e274d003a4d08fb888f1520add4b21373" ], "author": { "name": "Peter Zijlstra", "email": "a.p.zijlstra@chello.nl", "time": "Tue Oct 16 23:25:46 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Oct 17 08:42:45 2007 -0700" }, "message": "mm: bdi init hooks\n\nprovide BDI constructor/destructor hooks\n\n[akpm@linux-foundation.org: compile fix]\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6d66f5cd26e4c482e986130b7572f2735a0f7e8b", "tree": "b35a4c00af8b2120a672dc19cb09006d5d000a9f", "parents": [ "f0e7e1bd77d450ebfa12153b90f93ad46616ab4a" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 17:31:38 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:12 2007 -0700" }, "message": "sysfs: add copyrights\n\nSysfs has gone through considerable amount of reimplementation. Add\ncopyrights. Any objections? :-)\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "a4e8b912541d5372ae049a3b7c1979968e52c40b", "tree": "a95937748b6216f914abe55ed47c2627e7856bf7", "parents": [ "85a4ffad3de77177591f7c2c18c26c3c8dd28bff" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:12 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:11 2007 -0700" }, "message": "sysfs: move sysfs file poll implementation to sysfs_open_dirent\n\nSysfs file poll implementation is scattered over sysfs and kobject.\nEvent numbering is done in sysfs_dirent but wait itself is done on\nkobject. This not only unecessarily bloats both kobject and\nsysfs_dirent but is also buggy - if a sysfs_dirent is removed while\nthere still are pollers, the associaton betwen the kobject and\nsysfs_dirent breaks and kobject may be freed with the pollers still\nsleeping on it.\n\nThis patch moves whole poll implementation into sysfs_open_dirent.\nEach time a sysfs_open_dirent is created, event number restarts from 1\nand pollers sleep on sysfs_open_dirent. As event sequence number is\nmeaningless without any open file and pollers should have open file\nand thus sysfs_open_dirent, this ephemeral event counting works and is\na saner implementation.\n\nThis patch fixes the dnagling sleepers bug and reduces the sizes of\nkobject and sysfs_dirent by one pointer.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "85a4ffad3de77177591f7c2c18c26c3c8dd28bff", "tree": "3c67b1435dfd9a5b56d39235f1b44769f96f0fcc", "parents": [ "bc747f37a0f089b9366f7385ff870e12911f2383" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:12 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:11 2007 -0700" }, "message": "sysfs: implement sysfs_open_dirent\n\nImplement sysfs_open_dirent which represents an open file (attribute)\nsysfs_dirent. A file sysfs_dirent with one or more open files have\none sysfs_dirent and all sysfs_buffers (one for each open instance)\nare linked to it.\n\nsysfs_open_dirent doesn\u0027t actually do anything yet but will be used to\noff-load things which are specific for open file sysfs_dirent from it.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "bc747f37a0f089b9366f7385ff870e12911f2383", "tree": "b23e81c5136486bc081770dd83595e658437ace8", "parents": [ "dc2f75f0e0cac645c22c85bcf429683b3fa0d2d9" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:12 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:11 2007 -0700" }, "message": "sysfs: move sysfs_dirent-\u003es_children into sysfs_dirent-\u003es_dir\n\nChildren list head is only meaninful for directory nodes. Move it\ninto s_dir. This doesn\u0027t save any space currently but it will with\nfurther changes.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "dc2f75f0e0cac645c22c85bcf429683b3fa0d2d9", "tree": "775a89f5abae0cdc9688a32a6209d74c61cb5975", "parents": [ "d6b4fd2faeb9ddf55ce09cf90b88981e579ee010" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:12 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:11 2007 -0700" }, "message": "sysfs: make sysfs_root a regular directory dirent\n\nsysfs_root is different from a regular directory dirent in that it\u0027s\nof type SYSFS_ROOT and doesn\u0027t have a name. These differences aren\u0027t\nused by anybody and only adds to complexity. Make sysfs_root a\nregular directory dirent.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "b1fc3d6144d56360d1373b01c7881826f558b6cd", "tree": "3bac4c3df21dc74c6373c3e4a22c3ea6b01b1832", "parents": [ "078ce6409ca54d5fc6eb7d2147cd6efc3eb09078" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:11 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:10 2007 -0700" }, "message": "sysfs: make s_elem an anonymous union\n\nMake s_elem an anonymous union. Prefixing with s_elem makes things\nneedlessly longer without any advantage.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "b13dc89c5a5bd5e34aadb44c0fb7e870959dcd06", "tree": "110bf837aa041960726dc4ca592b7469907de3d8", "parents": [ "5a7ad7f044941316dc98eda2a087a12a7a50649d" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:10 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:10 2007 -0700" }, "message": "sysfs: reposition sysfs_dirent-\u003es_mode.\n\nMove s_mode downward such that it\u0027s side-by-side with s_iattr which is\nused for the same thing.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "59f69015684b3de7b9472be9a81b1a978f93a496", "tree": "3d3f7880e9f0d9c867695b120944540e815f194a", "parents": [ "f88123eaf953f13a0c597dde54745d28f81236de" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Sep 20 16:05:10 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:09 2007 -0700" }, "message": "sysfs: clean up header files\n\nsysfs is about to go through major overhaul making this a pretty good\nopportunity to clean up (out-of-tree changes and pending patches will\nneed regeneration anyway). Clean up headers.\n\n* Kill space between * and symbolname.\n\n* Move SYSFS_* type constants and flags into fs/sysfs/sysfs.h.\n They\u0027re internal to sysfs.\n\n* Reformat function prototypes and add argument symbol names.\n\n* Make dummy function definition order match that of function\n prototypes.\n\n* Add some comments.\n\n* Reorganize fs/sysfs/sysfs.h according to which file the declared\n variable or feature lives in.\n\nThis patch does not introduce any behavior change.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "5a26b79c426f8e55ebf7204cb138eb6b1645d4d3", "tree": "4cdcd80f88a479a6ed1ef53042f3b17359c0e80e", "parents": [ "e0712bbfd9cb617fc3a822781c2466fb6b7ede50" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Mon Aug 20 21:36:30 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:08 2007 -0700" }, "message": "sysfs: Remove s_dentry\n\nThe only uses of s_dentry left are the code that maintains\ns_dentry and trivial users that don\u0027t actually need it.\nSo this patch removes the s_dentry maintenance code and\nrestructures the trivial uses to use something else.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nCc: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "932ea2e374dd1ca26676297a5eccd1cdab86f7cd", "tree": "83bba794de7f1f9f9290192f4f95a8ed9c91dfb6", "parents": [ "89bec09705d2033b8b765f3c3ac5093f80bd5bc4" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Mon Aug 20 21:36:30 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:08 2007 -0700" }, "message": "sysfs: Introduce sysfs_rename_mutex\n\nLooking carefully at the rename code we have a subtle dependency\nthat the structure of sysfs not change while we are performing\na rename. If the parent directory of the object we are renaming\nchanges while the rename is being performed nasty things could\nhappen when we go to release our locks.\n\nSo introduce a sysfs_rename_mutex to prevent this highly\nunlikely theoretical issue.\n\nIn addition hold sysfs_rename_mutex over all calls to\nsysfs_get_dentry. Allowing sysfs_get_dentry to be simplified\nin the future.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nCc: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "7d0c7d676cc066413e1583b5af9fba8011972d41", "tree": "0afe89f9eca5e1e0ff31f5409a3568991cfceda2", "parents": [ "0333cd8a3f4249fde2c50929a6eac35245fc685b" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Mon Aug 20 21:36:30 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:07 2007 -0700" }, "message": "sysfs: Make sysfs_mount static\n\nThis patch modifies the users of sysfs_mount to use sysfs_root\ninstead (which is what they are looking for). It then\nmakes sysfs_mount static to keep people from using it\nby accident.\n\nThe net result is slightly faster and cleaner code.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nCc: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "119dd52be33dfe6285f586ab7354897fdefc7e23", "tree": "f7422f753f3605401d4ee0a34b85fb7a0038fbf3", "parents": [ "372e88bd1922228e0a55228f6dc8e311d1696fa0" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Mon Aug 20 21:36:29 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:07 2007 -0700" }, "message": "sysfs: Remove sysfs_instantiate\n\nNow that sysfs_get_inode is dropping the inode lock\nwe no longer have a need from sysfs_instantiate.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nCc: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "990e53f880be9ff93072b4cce590ec2826cee0b6", "tree": "faa7cbedef2f1d9b8d27832ccbaa28e508d134e8", "parents": [ "23dc279950a056c33a14d09cf759f5173d41abd9" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Aug 02 21:38:03 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:04 2007 -0700" }, "message": "sysfs: make sysfs_addrm_finish() return void\n\nWith the previous sysfs_add_one() update, there is only one user of\nthe return value of sysfs_addrm_finish() and the user can switch to\ntesting @sd easily. Make sysfs_addrm_finish() return void for cleaner\nsemantics as suggested by Satyam Sharma.\n\nThis patch doesn\u0027t introduce any noticeable behavior change.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nCc: Satyam Sharma \u003csatyam.sharma@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "23dc279950a056c33a14d09cf759f5173d41abd9", "tree": "852f8cc24c24988717dbae8f98da999ff27755fa", "parents": [ "41fc1c27452e041a18e5141b8203ee0ea72bc483" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Aug 02 21:38:03 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:04 2007 -0700" }, "message": "sysfs: make sysfs_add_one() automatically check for duplicate entry\n\nMake sysfs_add_one() check for duplicate entry and return -EEXIST if\nsuch entry exists. This simplifies node addition code a bit.\n\nThis patch doesn\u0027t introduce any noticeable behavior change.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "41fc1c27452e041a18e5141b8203ee0ea72bc483", "tree": "fc81138ca37182c25931818681630c90b018b127", "parents": [ "ff869de7bf5e76adffebd3a176c1c73bca7eddb7" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Aug 02 21:38:03 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:03 2007 -0700" }, "message": "sysfs: make sysfs_add/remove_one() call link/unlink_sibling() implictly\n\nWhen adding or removing a sysfs_dirent, the user used to be required\nto call link/unlink separately. It was for two reasons - code looked\nlike that before sysfs_addrm_cxt conversion and to avoid looping\nthrough parent_sd-\u003echildren list twice during removal.\n\nPerformance optimization during removal just isn\u0027t worth it. Make\nsysfs_add/remove_one() call sysfs_link/unlink_sibing() implicitly.\nThis makes code simpler albeit slightly less efficient. This change\ndoesn\u0027t introduce any noticeable behavior change.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "90bc61359de0148f8627073d68a22edc7ed9893d", "tree": "b054bf0cb9bda41dab498086216f4c0253b2c5ed", "parents": [ "869512ab5ab93e5e82ad7d4aaf4ed098d23bfc3f" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Tue Jul 31 19:15:08 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Oct 12 14:51:03 2007 -0700" }, "message": "sysfs: Remove first pass at shadow directory support\n\nWhile shadow directories appear to be a good idea, the current scheme\nof controlling their creation and destruction outside of sysfs appears\nto be a locking and maintenance nightmare in the face of sysfs\ndirectories dynamically coming and going. Which can now occur for\ndirectories containing network devices when CONFIG_SYSFS_DEPRECATED is\nnot set.\n\nThis patch removes everything from the initial shadow directory support\nthat allowed the shadow directory creation to be controlled at a higher\nlevel. So except for a few bits of sysfs_rename_dir everything from\ncommit b592fcfe7f06c15ec11774b5be7ce0de3aa86e73 is now gone.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "bc37e2830339cbfa42ac8579a7bf62fc4cdd360d", "tree": "d01ee3243073407ec83c217402273976000f0041", "parents": [ "e080e436f605877e47e4950f5386ed843badbb1b" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Wed Jul 18 14:30:28 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 18 15:49:49 2007 -0700" }, "message": "sysfs: make sysfs_init_inode() static\n\nWith sysfs_fill_super() converted to use sysfs_get_inode(), there is\nno user of sysfs_init_inode() outside of fs/sysfs/inode.c. Make it\nstatic.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nAcked-by: Jean Delvare \u003ckhali@linux-fr.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "51225039f3cf9d250596d1344494b293274b9169", "tree": "4e646ab37043068824a146c7353c6242a4eb72df", "parents": [ "53e0ae92690c52eceb997905d85fbb42de5fff63" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:25 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:09 2007 -0700" }, "message": "sysfs: make directory dentries and inodes reclaimable\n\nThis patch makes dentries and inodes for sysfs directories\nreclaimable.\n\n* sysfs_notify() is modified to walk sysfs_dirent tree instead of\n dentry tree.\n\n* sysfs_update_file() and sysfs_chmod_file() use sysfs_get_dentry() to\n grab the victim dentry.\n\n* sysfs_rename_dir() and sysfs_move_dir() grab all dentries using\n sysfs_get_dentry() on startup.\n\n* Dentries for all shadowed directories are pinned in memory to serve\n as lookup start point.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "53e0ae92690c52eceb997905d85fbb42de5fff63", "tree": "b3bccd211bb09d43c29a751793ac50d8860b8303", "parents": [ "a0edd7c848945a75e2f41673f43bc37d0a5fed15" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:25 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:09 2007 -0700" }, "message": "sysfs: implement sysfs_get_dentry()\n\nSome sysfs operations require dentry and inode. sysfs_get_dentry()\nlooks up and gets dentry for the specified sysfs_dirent. It finds the\nfirst ancestor with dentry attached and starts looking up dentries\nfrom there.\n\nLooking up from the nearest ancestor is necessary to support shadowed\ndirectories because we can\u0027t reliably lookup dentry for one of the\nshadows. Dentries for each shadow will be pinned in memory such that\nthey can serve as the starting point for dentry lookup.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "a0edd7c848945a75e2f41673f43bc37d0a5fed15", "tree": "c74ea7d7dbc814faef126f97601a16c7118b3099", "parents": [ "fb6896da37f19be4b75154c14d1cd79231255b17" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:24 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:09 2007 -0700" }, "message": "sysfs: move sysfs_drop_dentry() to dir.c and make it static\n\nAfter add/remove path restructuring, the only user of\nsysfs_drop_dentry() is sysfs_addrm_finish(). Move sysfs_drop_dentry()\nto dir.c and make it static.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "fb6896da37f19be4b75154c14d1cd79231255b17", "tree": "d1480206694a0642067dc7dc5b40bd63b038cac0", "parents": [ "3007e997de91ec59af39a3f9c91595b31ae6e08b" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:24 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:09 2007 -0700" }, "message": "sysfs: restructure add/remove paths and fix inode update\n\nThe original add/remove code had the following problems.\n\n* parent\u0027s timestamps are updated on dentry instantiation. this is\n incorrect with reclaimable files.\n\n* updating parent\u0027s timestamps isn\u0027t synchronized.\n\n* parent nlink update assumes the inode is accessible which won\u0027t be\n true once directory dentries are made reclaimable.\n\nThis patch restructures add/remove paths to resolve the above\nproblems. Add/removal are done in the following steps.\n\n1. sysfs_addrm_start() : acquire locks including sysfs_mutex and other\n resources.\n\n2-a. sysfs_add_one() : add new sd. linking the new sd into the\n children list is caller\u0027s responsibility.\n\n2-b. sysfs_remove_one() : remove a sd. unlinking the sd from the\n children list is caller\u0027s responsibility.\n\n3. sysfs_addrm_finish() : release all resources and clean up.\n\nSteps 2-a and/or 2-b can be repeated multiple times.\n\nParent\u0027s inode is looked up during sysfs_addrm_start(). If available\n(always at the moment), it\u0027s pinned and nlink is updated as sd\u0027s are\nadded and removed. Timestamps are updated during finish if any sd has\nbeen added or removed. If parent\u0027s inode is not available during\nstart, sysfs_mutex ensures that parent inode is not created till\nadd/remove is complete.\n\nAll the complexity is contained inside the helper functions.\nEspecially, dentry/inode handling is properly hidden from the rest of\nsysfs which now mostly operate on sysfs_dirents. As an added bonus,\ncodes which use these helpers to add and remove sysfs_dirents are now\nmore structured and simpler.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "3007e997de91ec59af39a3f9c91595b31ae6e08b", "tree": "4ed4df3ef3a249d2a4b562e36876fc8d4a3fabd9", "parents": [ "5f9953237f684ea1778adb9d26162da00b282225" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:23 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:08 2007 -0700" }, "message": "sysfs: use sysfs_mutex to protect the sysfs_dirent tree\n\nAs kobj sysfs dentries and inodes are gonna be made reclaimable,\ni_mutex can\u0027t be used to protect sysfs_dirent tree. Use sysfs_mutex\nglobally instead. As the whole tree is protected with sysfs_mutex,\nthere is no reason to keep sysfs_rename_sem. Drop it.\n\nWhile at it, add docbook comments to functions which require\nsysfs_mutex locking.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "5f9953237f684ea1778adb9d26162da00b282225", "tree": "962bbe5fe2483876caa74943875cbef432db6945", "parents": [ "608e266a2d4e62c1b98c1c573064b6afe8c06a58" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:23 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:08 2007 -0700" }, "message": "sysfs: consolidate sysfs spinlocks\n\nReplace sysfs_lock and kobj_sysfs_assoc_lock with sysfs_assoc_lock.\nsysfs_lock was originally to be used to protect sysfs_dirent tree but\nmutex seems better choice, so there is no reason to keep sysfs_lock\nseparate. Merge the two spinlocks into one.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "608e266a2d4e62c1b98c1c573064b6afe8c06a58", "tree": "73d670ab6c315b42273bc9d2731ec2b8e7c199af", "parents": [ "f0b0af4792d751106e2003f96af76fa95e10c68d" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:22 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:08 2007 -0700" }, "message": "sysfs: make kobj point to sysfs_dirent instead of dentry\n\nAs kobj sysfs dentries and inodes are gonna be made reclaimable,\ndentry can\u0027t be used as naming token for sysfs file/directory, replace\nkobj-\u003edentry with kobj-\u003esd. The only external interface change is\nshadow directory handling. All other changes are contained in kobj\nand sysfs.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "f0b0af4792d751106e2003f96af76fa95e10c68d", "tree": "09483c5b34009cc815897d885f310d0704d0c396", "parents": [ "380e6fbb729a55b73d5d8409551474884e0d93fc" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:22 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:08 2007 -0700" }, "message": "sysfs: implement sysfs_find_dirent() and sysfs_get_dirent()\n\nImplement sysfs_find_dirent() and sysfs_get_dirent().\nsysfs_dirent_exist() is replaced by sysfs_find_dirent(). These will\nbe used to make directory entries reclamiable.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "b402d72cf7b338a074e3c12b305ec79284e18845", "tree": "0724d40180eb78212be81c7ff2d909ec3fb3f306", "parents": [ "d0bcb5689a521df98bff7549fcb8b17499660a99" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 04:27:21 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:08 2007 -0700" }, "message": "sysfs: rename sysfs_dirent-\u003es_type to s_flags and make room for flags\n\nRename sysfs_dirent-\u003es_type to s_flags, pack type into lower eight\nbits and reserve the rest for flags. sysfs_type() can used to access\nthe type. All existing sd-\u003es_type accesses are converted to use\nsysfs_type(). While at it, type test is changed to equality test\ninstead of bit-and test where appropriate.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "0c73f18b7d95de8a007039337063a770b5fc8e7a", "tree": "cf3b6f274139b170136219918dd679f2397e69f6", "parents": [ "8619f979898397582e366877fd5feeba7560d70c" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:18 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:07 2007 -0700" }, "message": "sysfs: use singly-linked list for sysfs_dirent tree\n\nMake sysfs_dirent use singly linked list for its tree structure.\nsysfs_link_sibling() and sysfs_unlink_sibling() functions are added to\nhandle simpler cases. It adds some complexity and cpu cycle overhead\nbut reduced memory footprint is worthwhile on big machines.\n\nThis change reduces the sizeof sysfs_dirent from 104 to 88 on 64bit\nand from 60 to 52 on 32bit.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "8619f979898397582e366877fd5feeba7560d70c", "tree": "7d0e92d4c17d92a00dfeba2b840bf89657b30607", "parents": [ "b6b4a4399c2a83d1af77c99dee0d0b5cc15ec268" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:18 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:07 2007 -0700" }, "message": "sysfs: slim down sysfs_dirent-\u003es_active\n\nMake sysfs_dirent-\u003es_active an atomic_t instead of rwsem. This\nreduces the size of sysfs_dirent from 136 to 104 on 64bit and from 76\nto 60 on 32bit with lock debugging turned off. With lock debugging\nturned on the reduction is much larger.\n\ns_active starts at zero and each active reference increments s_active.\nPutting a reference decrements s_active. Deactivation subtracts\nSD_DEACTIVATED_BIAS which is currently INT_MIN and assumed to be small\nenough to make s_active negative. If s_active is negative,\nsysfs_get() no longer grants new references. Deactivation succeeds\nimmediately if there is no active user; otherwise, it waits using a\ncompletion for the last put.\n\nDue to the removal of lockdep tricks, this change makes things less\ntrickier in release_sysfs_dirent(). As all the complexity is\ncontained in three s_active functions, I think it\u0027s more readable this\nway.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "b6b4a4399c2a83d1af77c99dee0d0b5cc15ec268", "tree": "7a5d3134a338f03ab813024d5501bde1d434c98b", "parents": [ "0b8ead82f5d9d8f08c0d1236f2e350b70a977753" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:18 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:07 2007 -0700" }, "message": "sysfs: move s_active functions to fs/sysfs/dir.c\n\nThese functions are about to receive more complexity and doesn\u0027t\nreally need to be inlined in the first place. Move them from\nfs/sysfs/sysfs.h to fs/sysfs/dir.c.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "8312a8d7c1d19d31027bd4ca127ce671962c23d4", "tree": "ee28a44611ac2192b265f5f85a74f7da98d2fdef", "parents": [ "fc9f54b9982e14e6dbe023425c87ffbfd6992c45" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:17 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:07 2007 -0700" }, "message": "sysfs: use iget_locked() instead of new_inode()\n\nAfter dentry is reclaimed, sysfs always used to allocate new dentry\nand inode if the file is accessed again. This causes problem with\noperations which only pin the inode. For example, if inotify watch is\nadded to a sysfs file and the dentry for the file is reclaimed, the\nnext update event creates new dentry and new inode making the inotify\nwatch miss all the events from there on.\n\nThis patch fixes it by using iget_locked() instead of new_inode().\nsysfs_new_inode() is renamed to sysfs_get_inode() and inode is\ninitialized iff the inode is newly allocated. sysfs_instantiate() is\nresponsible for unlocking new inodes.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "fc9f54b9982e14e6dbe023425c87ffbfd6992c45", "tree": "61b828ac694d9e8decb02022a1056a5367bb1bef", "parents": [ "7f7cfffe60ed6271c4028ec79ae1c297b44bcb14" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:17 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:07 2007 -0700" }, "message": "sysfs: reorganize sysfs_new_indoe() and sysfs_create()\n\nReorganize/clean up sysfs_new_inode() and sysfs_create().\n\n* sysfs_init_inode() is separated out from sysfs_new_inode() and is\n responsible for basic initialization.\n* sysfs_instantiate() replaces the last step of sysfs_create() and is\n responsible for dentry instantitaion.\n* type-specific initialization is moved out to the callers.\n* mode is specified only once when creating a sysfs_dirent.\n* spurious list_del_init(\u0026sd-\u003es_sibling) dropped from create_dir()\n\nThis change is to\n\n* prepare for inode allocation fix.\n* separate alloc and init code for synchronization update.\n* make dentry/inode initialization more flexible for later changes.\n\nThis patch doesn\u0027t introduce visible behavior change.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "dbde0fcf9f8f6d477af3c32d9979e789ee680cde", "tree": "12a1240213d59c061fec60325c7d5ebb1edebcd7", "parents": [ "198a2a847015805c6f57d8cc732bdaaccb494007" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:16 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:06 2007 -0700" }, "message": "sysfs: reimplement sysfs_drop_dentry()\n\nThis patch reimplements sysfs_drop_dentry() such that remove_dir() can\nuse it to drop dentry instead of using a separate mechanism. With\nthis change, making directories reclaimable is much easier.\n\nThis patch used to contain fixes for two race conditions around\nsd-\u003es_dentry but that part has been separated out and included into\nmainline early as commit 6aa054aadfea613a437ad0b15d38eca2b963fc0a and\ndd14cbc994709a1c5a64ed3621f583c49a27e521.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "198a2a847015805c6f57d8cc732bdaaccb494007", "tree": "614aa44e1bd1ab31a5baaab717c3af95773bd06b", "parents": [ "73107cb3ad3963c0f929ae681c05081eafb1c079" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:16 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:05 2007 -0700" }, "message": "sysfs: separate out sysfs_attach_dentry()\n\nConsolidate sd \u003c-\u003e dentry association into sysfs_attach_dentry() and\ncall it after dentry and inode are properly set up. This is in\npreparation of sysfs_drop_dentry() updates.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "73107cb3ad3963c0f929ae681c05081eafb1c079", "tree": "78c9647d5e6a0d9c89cb8c37b7533aad8f805302", "parents": [ "0ab66088c855eca68513bdd7442a426c4b374ced" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:16 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:05 2007 -0700" }, "message": "sysfs: kill attribute file orphaning\n\nNow that sysfs_dirent can be disconnected from kobject on deletion,\nthere is no need to orphan each attribute files. All [bin_]attribute\nnodes are automatically orphaned when the parent node is deleted.\nKill attribute file orphaning.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "0ab66088c855eca68513bdd7442a426c4b374ced", "tree": "7f931f7f984ac14701a33d123fa9e03d9048bf8f", "parents": [ "eb36165353d0e5ac32b063f555acedcbaf6d3b75" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:16 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:05 2007 -0700" }, "message": "sysfs: implement sysfs_dirent active reference and immediate disconnect\n\nsysfs: implement sysfs_dirent active reference and immediate disconnect\n\nOpening a sysfs node references its associated kobject, so userland\ncan arbitrarily prolong lifetime of a kobject which complicates\nlifetime rules in drivers. This patch implements active reference and\nmakes the association between kobject and sysfs immediately breakable.\n\nNow each sysfs_dirent has two reference counts - s_count and s_active.\ns_count is a regular reference count which guarantees that the\ncontaining sysfs_dirent is accessible. As long as s_count reference\nis held, all sysfs internal fields in sysfs_dirent are accessible\nincluding s_parent and s_name.\n\nThe newly added s_active is active reference count. This is acquired\nby invoking sysfs_get_active() and it\u0027s the caller\u0027s responsibility to\nensure sysfs_dirent itself is accessible (should be holding s_count\none way or the other). Dereferencing sysfs_dirent to access objects\nout of sysfs proper requires active reference. This includes access\nto the associated kobjects, attributes and ops.\n\nThe active references can be drained and denied by calling\nsysfs_deactivate(). All active sysfs_dirents must be deactivated\nafter deletion but before the default reference is dropped. This\nenables immediate disconnect of sysfs nodes. Once a sysfs_dirent is\ndeleted, it won\u0027t access any entity external to sysfs proper.\n\nBecause attr/bin_attr ops access both the node itself and its parent\nfor kobject, they need to hold active references to both.\nsysfs_get/put_active_two() helpers are provided to help grabbing both\nreferences. Parent\u0027s is acquired first and released last.\n\nUnlike other operations, mmapped area lingers on after mmap() is\nfinished and the module implement implementing it and kobj need to\nstay referenced till all the mapped pages are gone. This is\naccomplished by holding one set of active references to the bin_attr\nand its parent if there have been any mmap during lifetime of an\nopenfile. The references are dropped when the openfile is released.\n\nThis change makes sysfs lifetime rules independent from both kobject\u0027s\nand module\u0027s. It not only fixes several race conditions caused by\nsysfs not holding onto the proper module when referencing kobject, but\nalso helps fixing and simplifying lifetime management in driver model\nand drivers by taking sysfs out of the equation.\n\nPlease read the following message for more info.\n\n http://article.gmane.org/gmane.linux.kernel/510293\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "2b29ac252afff87b8465b064ca2d9740cf1f6e52", "tree": "9f4930db68ace50adc7c11feba12aafe34e2cdbe", "parents": [ "aecdcedaab49ca40620dc7dd70f67ee7269a66c9" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:15 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:04 2007 -0700" }, "message": "sysfs: reimplement symlink using sysfs_dirent tree\n\nsysfs symlink is implemented by referencing dentry and kobject from\nsysfs_dirent - symlink entry references kobject, dentry is used to\nwalk the tree. This complicates object lifetimes rules and is\ndangerous - for example, there is no way to tell to which module the\ntarget of a symlink belongs and referencing that kobject can make it\nlinger after the module is gone.\n\nThis patch reimplements symlink using only sysfs_dirent tree. sd for\na symlink points and holds reference to the target sysfs_dirent and\nall walking is done using sysfs_dirent tree. Simpler and safer.\n\nPlease read the following message for more info.\n\n http://article.gmane.org/gmane.linux.kernel/510293\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "aecdcedaab49ca40620dc7dd70f67ee7269a66c9", "tree": "df1faa0302e6fb428b7ae0a14d443626fd73b121", "parents": [ "3e5190380ebef77f2b015c9e7a4ca225a3d75021" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:15 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:04 2007 -0700" }, "message": "sysfs: implement kobj_sysfs_assoc_lock\n\nkobj-\u003edentry can go away anytime unless the user controls when the\nassociated sysfs node is deleted. This patch implements\nkobj_sysfs_assoc_lock which protects kobj-\u003edentry. This will be used\nto maintain kobj based API when converting sysfs to use sysfs_dirent\ntree instead of dentry/kobject.\n\nNote that this lock belongs to kobject/driver-model not sysfs. Once\nsysfs is converted to not use kobject in its interface, this can be\nremoved from sysfs.\n\nThis is in preparation of object reference simplification.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "3e5190380ebef77f2b015c9e7a4ca225a3d75021", "tree": "08d17cd0562acf5101063e93c649f5f71f80b3b9", "parents": [ "0c096b507f15397da890051ee73de4266d3941fb" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:15 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:04 2007 -0700" }, "message": "sysfs: make sysfs_dirent-\u003es_element a union\n\nMake sd-\u003es_element a union of sysfs_elem_{dir|symlink|attr|bin_attr}\nand rename it to s_elem. This is to achieve...\n\n* some level of type checking : changing symlink to point to\n sysfs_dirent instead of kobject is much safer and less painful now.\n* easier / standardized dereferencing\n* allow sysfs_elem_* to contain more than one entry\n\nWhere possible, pointer is obtained by directly deferencing from sd\ninstead of going through other entities. This reduces dependencies to\ndentry, inode and kobject. to_attr() and to_bin_attr() are unused now\nand removed.\n\nThis is in preparation of object reference simplification.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "0c096b507f15397da890051ee73de4266d3941fb", "tree": "c6b7d204a493157b3926bdb46500313eee9f9690", "parents": [ "13b3086d2ea483cbcae5a4236446cecc082a72cf" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:15 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:04 2007 -0700" }, "message": "sysfs: add sysfs_dirent-\u003es_name\n\nAdd s_name to sysfs_dirent. This is to further reduce dependency to\nthe associated dentry. Name is copied for directories and symlinks\nbut not for attributes.\n\nWhere possible, name dereferences are converted to use sd-\u003es_name.\nsysfs_symlink-\u003elink_name and sysfs_get_name() are unused now and\nremoved.\n\nThis change allows symlink to be implemented using sysfs_dirent tree\nproper, which is the last remaining dentry-dependent sysfs walk.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "13b3086d2ea483cbcae5a4236446cecc082a72cf", "tree": "f6932aaf486a302aa7b03857d780bafda33db2d6", "parents": [ "a26cd7226c24c3be5dd5f48a74832fe64beb8489" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:14 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:04 2007 -0700" }, "message": "sysfs: add sysfs_dirent-\u003es_parent\n\nAdd sysfs_dirent-\u003es_parent. With this patch, each sd points to and\nholds a reference to its parent. This allows walking sysfs tree\nwithout referencing sd-\u003es_dentry which can go away anytime if the user\ndoesn\u0027t control when it\u0027s deleted.\n\nsd-\u003es_parent is initialized and parent is referenced in\nsysfs_attach_dirent(). Reference to parent is released when the sd is\nreleased, so as long as reference to a sd is held, s_parent can be\nfollowed.\n\ndentry walk in sysfs_readdir() is convereted to s_parent walk.\n\nThis will be used to reimplement symlink such that it uses only\nsysfs_dirent tree.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "a26cd7226c24c3be5dd5f48a74832fe64beb8489", "tree": "0dfb4a15d800494a06a4ddf83c4d33812c404fac", "parents": [ "996b73764e9bb9d5e751fd15b130ba38637d66a8" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:14 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:04 2007 -0700" }, "message": "sysfs: consolidate sysfs_dirent creation functions\n\nCurrently there are four functions to create sysfs_dirent -\n__sysfs_new_dirent(), sysfs_new_dirent(), __sysfs_make_dirent() and\nsysfs_make_dirent(). Other than sysfs_make_dirent(), no function has\ntwo users if calls to implement other functions are excluded.\n\nThis patch consolidates sysfs_dirent creation functions into the\nfollowing two.\n\n* sysfs_new_dirent() : allocate and initialize\n* sysfs_attach_dirent() : attach to sysfs_dirent hierarchy and/or\n\t\t\t associate with dentry\n\nThis simplifies interface and gives callers more flexibility. This is\nin preparation of object reference simplification.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "7a23ad44047b1084a032bc0d127fe08af024593a", "tree": "1f1819b0897fdd8e0237f22d8eebeb55b75894b9", "parents": [ "2b611bb7abdcc08278453fc9f6517401fd69ef95" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:13 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:03 2007 -0700" }, "message": "sysfs: make sysfs_put() ignore NULL sd\n\nMake sysfs_put() ignore NULL sd instead of oopsing.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "fa7f912ad4ae0ed7591add52422e48282389652d", "tree": "2f986dbf55002e32d6d296a1b5b96d55135ad28f", "parents": [ "72dba584b695d8bc8c1a50ed54ad4cba7c62314d" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:13 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:03 2007 -0700" }, "message": "sysfs: move release_sysfs_dirent() to dir.c\n\nThere is no reason this function should be inlined and soon to follow\nsysfs object reference simplification will make it heavier. Move it\nto dir.c.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "dd14cbc994709a1c5a64ed3621f583c49a27e521", "tree": "e48d38b7450661907c7b75490504c7f70b04d6cc", "parents": [ "6aa054aadfea613a437ad0b15d38eca2b963fc0a" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Mon Jun 11 14:04:01 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Tue Jun 12 16:08:47 2007 -0700" }, "message": "sysfs: fix race condition around sd-\u003es_dentry, take#2\n\nAllowing attribute and symlink dentries to be reclaimed means\nsd-\u003es_dentry can change dynamically. However, updates to the field\nare unsynchronized leading to race conditions. This patch adds\nsysfs_lock and use it to synchronize updates to sd-\u003es_dentry.\n\nDue to the locking around -\u003ed_iput, the check in sysfs_drop_dentry()\nis complex. sysfs_lock only protect sd-\u003es_dentry pointer itself. The\nvalidity of the dentry is protected by dcache_lock, so whether dentry\nis alive or not can only be tested while holding both locks.\n\nThis is minimal backport of sysfs_drop_dentry() rewrite in devel\nbranch.\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "dc351252b33f8fede396d6173dba117bcb933607", "tree": "282d57855f66119f930eb629ab483bffcc5b6c21", "parents": [ "99f9f3d49cbc7d944476f6fde53a77ec789ab2aa" ], "author": { "name": "Eric Sandeen", "email": "sandeen@sandeen.net", "time": "Mon Jun 11 14:02:45 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Tue Jun 12 16:08:46 2007 -0700" }, "message": "sysfs: store sysfs inode nrs in s_ino to avoid readdir oopses\n\nBackport of\nftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.22-rc1/2.6.22-rc1-mm1/broken-out/gregkh-driver-sysfs-allocate-inode-number-using-ida.patch\n\nFor regular files in sysfs, sysfs_readdir wants to traverse\nsysfs_dirent-\u003es_dentry-\u003ed_inode-\u003ei_ino to get to the inode number.\nBut, the dentry can be reclaimed under memory pressure, and there is\nno synchronization with readdir. This patch follows Tejun\u0027s scheme of\nallocating and storing an inode number in the new s_ino member of a\nsysfs_dirent, when dirents are created, and retrieving it from there\nfor readdir, so that the pointer chain doesn\u0027t have to be traversed.\n\nTejun\u0027s upstream patch uses a new-ish \"ida\" allocator which brings\nalong some extra complexity; this -stable patch has a brain-dead\nincrementing counter which does not guarantee uniqueness, but because\nsysfs doesn\u0027t hash inodes as iunique expects, uniqueness wasn\u0027t\nguaranteed today anyway.\n\nSigned-off-by: Eric Sandeen \u003csandeen@redhat.com\u003e\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "d56c3eae6753f53d3313b926dcdda38c6c9bbe9b", "tree": "c5df874dec463be02221a58a9dcdaf43d8038457", "parents": [ "63ce18cfe685115ff8d341bae4c9204a79043cf0" ], "author": { "name": "Adam J. Richter", "email": "adam@yggdrasil.com", "time": "Fri Feb 16 21:35:25 2007 +0800" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Feb 23 14:52:09 2007 -0800" }, "message": "sysfs: move struct sysfs_dirent to private header\n\nstruct sysfs_dirent is private to the fs/sysfs/ subtree. It is\nnot even referenced as an opaque structure outside of that subtree.\n\nThe following patch moves the declaration from include/linux/sysfs.h to\nfs/sysfs/sysfs.h, making it clearer that nothing else in the kernel\ndereferences it.\n\nI have been running this patch for years. Please integrate and forward\nupstream if there are no objections.\n\nFrom: \"Adam J. Richter\" \u003cadam@yggdrasil.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "c5ef1c42c51b1b5b4a401a6517bdda30933ddbaf", "tree": "e1a9804a8af427f700aaba4b386cf8679b317e83", "parents": [ "92e1d5be91a0e3ffa5c4697eeb09b2aa22792122" ], "author": { "name": "Arjan van de Ven", "email": "arjan@linux.intel.com", "time": "Mon Feb 12 00:55:40 2007 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Mon Feb 12 09:48:46 2007 -0800" }, "message": "[PATCH] mark struct inode_operations const 3\n\nMany struct inode_operations in the kernel can be \"const\". Marking them const\nmoves these to the .rodata section, which avoids false sharing with potential\ndirty data. In addition it\u0027ll catch accidental writes at compile time to\nthese shared resources.\n\nSigned-off-by: Arjan van de Ven \u003carjan@linux.intel.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b592fcfe7f06c15ec11774b5be7ce0de3aa86e73", "tree": "13f2cb344f8871edd30dc15007534405197d8480", "parents": [ "2f65168de7d68a5795e945e781d85b313bdc97b9" ], "author": { "name": "Eric W. Biederman", "email": "ebiederm@xmission.com", "time": "Wed Jan 24 12:35:52 2007 -0700" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Feb 07 10:37:14 2007 -0800" }, "message": "sysfs: Shadow directory support\n\nThe problem. When implementing a network namespace I need to be able\nto have multiple network devices with the same name. Currently this\nis a problem for /sys/class/net/*. \n\nWhat I want is a separate /sys/class/net directory in sysfs for each\nnetwork namespace, and I want to name each of them /sys/class/net.\n\nI looked and the VFS actually allows that. All that is needed is\nfor /sys/class/net to implement a follow link method to redirect\nlookups to the real directory you want. \n\nImplementing a follow link method that is sensitive to the current\nnetwork namespace turns out to be 3 lines of code so it looks like a\nclean approach. Modifying sysfs so it doesn\u0027t get in my was is a bit\ntrickier. \n\nI am calling the concept of multiple directories all at the same path\nin the filesystem shadow directories. With the directory entry really\nat that location the shadow master. \n\nThe following patch modifies sysfs so it can handle a directory\nstructure slightly different from the kobject tree so I can implement\nthe shadow directories for handling /sys/class/net/.\n\nSigned-off-by: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nCc: Maneesh Soni \u003cmaneesh@in.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "94bebf4d1b8e7719f0f3944c037a21cfd99a4af7", "tree": "25c6ba1836e74f608b4e434b4f0f4d5c28b11de1", "parents": [ "cb986b749c7178422bfbc982cd30e04d5db54bbc" ], "author": { "name": "Oliver Neukum", "email": "oliver@neukum.org", "time": "Wed Dec 20 10:52:44 2006 +0100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Feb 07 10:37:13 2007 -0800" }, "message": "Driver core: fix race in sysfs between sysfs_remove_file() and read()/write()\n\nThis patch prevents a race between IO and removing a file from sysfs.\nIt introduces a list of sysfs_buffers associated with a file at the inode.\nUpon removal of a file the list is walked and the buffers marked orphaned.\nIO to orphaned buffers fails with -ENODEV. The driver can safely free\nassociated data structures or be unloaded.\n\nSigned-off-by: Oliver Neukum \u003coliver@neukum.name\u003e\nAcked-by: Maneesh Soni \u003cmaneesh@in.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n\n" }, { "commit": "e18b890bb0881bbab6f4f1a6cd20d9c60d66b003", "tree": "4828be07e1c24781c264b42c5a75bcd968223c3f", "parents": [ "441e143e95f5aa1e04026cb0aa71c801ba53982f" ], "author": { "name": "Christoph Lameter", "email": "clameter@sgi.com", "time": "Wed Dec 06 20:33:20 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.osdl.org", "time": "Thu Dec 07 08:39:25 2006 -0800" }, "message": "[PATCH] slab: remove kmem_cache_t\n\nReplace all uses of kmem_cache_t with struct kmem_cache.\n\nThe patch was generated using the following script:\n\n\t#!/bin/sh\n\t#\n\t# Replace one string by another in all the kernel sources.\n\t#\n\n\tset -e\n\n\tfor file in `find * -name \"*.c\" -o -name \"*.h\"|xargs grep -l $1`; do\n\t\tquilt add $file\n\t\tsed -e \"1,\\$s/$1/$2/g\" $file \u003e/tmp/$$\n\t\tmv /tmp/$$ $file\n\t\tquilt refresh\n\tdone\n\nThe script was run like this\n\n\tsh replace kmem_cache_t \"struct kmem_cache\"\n\nSigned-off-by: Christoph Lameter \u003cclameter@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "995982ca79d9262869513948ec7c540f32035491", "tree": "08c9a655a210388dce467400aa11484cfd2b230b", "parents": [ "10188012daa586ae7fcbef272e4db4f404741adf" ], "author": { "name": "Randy.Dunlap", "email": "rdunlap@xenotime.net", "time": "Mon Jul 10 23:05:25 2006 -0700" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Mon Sep 25 21:08:39 2006 -0700" }, "message": "sysfs_remove_bin_file: no return value, dump_stack on error\n\nMake sysfs_remove_bin_file() void. If it detects an error,\nprintk the file name and call dump_stack().\n\nsysfs_hash_and_remove() now returns an error code indicating\nits success or failure so that sysfs_remove_bin_file() can\nknow success/failure.\n\nConvert the only driver that checked the return value of\nsysfs_remove_bin_file().\n\nSigned-off-by: Randy Dunlap \u003crdunlap@xenotime.net\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "4508a7a734b111b8b7e39986237d84acb1168dd0", "tree": "8fe535fbb97d619c9069da26367ead03d1a294aa", "parents": [ "f043ca43c1ae354346f72dc5826d820d5619f0b2" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 20 17:53:53 2006 +1100" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Fri Apr 14 11:41:24 2006 -0700" }, "message": "[PATCH] sysfs: Allow sysfs attribute files to be pollable\n\nIt works like this:\n Open the file\n Read all the contents.\n Call poll requesting POLLERR or POLLPRI (so select/exceptfds works)\n When poll returns,\n close the file and go to top of loop.\n or lseek to start of file and go back to the \u0027read\u0027.\n\nEvents are signaled by an object manager calling\n sysfs_notify(kobj, dir, attr);\n\nIf the dir is non-NULL, it is used to find a subdirectory which\ncontains the attribute (presumably created by sysfs_create_group).\n\nThis has a cost of one int per attribute, one wait_queuehead per kobject,\none int per open file.\n\nThe name \"sysfs_notify\" may be confused with the inotify\nfunctionality. Maybe it would be nice to support inotify for sysfs\nattributes as well?\n\nThis patch also uses sysfs_notify to allow /sys/block/md*/md/sync_action\nto be pollable\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n" }, { "commit": "4b6f5d20b04dcbc3d888555522b90ba6d36c4106", "tree": "420f271eaef7d3def7d4433b151c3cb6d7a54770", "parents": [ "99ac48f54a91d02140c497edc31dc57d4bc5c85d" ], "author": { "name": "Arjan van de Ven", "email": "arjan@infradead.org", "time": "Tue Mar 28 01:56:42 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Tue Mar 28 09:16:06 2006 -0800" }, "message": "[PATCH] Make most file operations structs in fs/ const\n\nThis is a conversion to make the various file_operations structs in fs/\nconst. Basically a regexp job, with a few manual fixups\n\nThe goal is both to increase correctness (harder to accidentally write to\nshared datastructures) and reducing the false sharing of cachelines with\nthings that get dirty in .data (while .rodata is nicely read only and thus\ncache clean)\n\nSigned-off-by: Arjan van de Ven \u003carjan@infradead.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "c516865cfbac0d862d4888df91793ad1e74ffd58", "tree": "432024125976af3e6c87ae5b9e64b6f1cc291f70", "parents": [ "22f98c0cd7e003b896ee52ded945081307118745" ], "author": { "name": "Maneesh Soni", "email": "maneesh@in.ibm.com", "time": "Thu Mar 09 19:40:14 2006 +0530" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Mon Mar 20 13:42:59 2006 -0800" }, "message": "[PATCH] sysfs: fix problem with duplicate sysfs directories and files\n\nThe following patch checks for existing sysfs_dirent before\npreparing new one while creating sysfs directories and files.\n\nSigned-off-by: Maneesh Soni \u003cmaneesh@in.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n" }, { "commit": "5f45f1a78fbac3cc859ec10c5366e97d20d40fa2", "tree": "bafacf11adbfb4eebf2d2d80d92e1fe2d9c02b18", "parents": [ "ac20427ef6aa63da663bdc88b71d16f7394f5e23" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Thu Jun 23 00:09:12 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Thu Jun 23 09:45:20 2005 -0700" }, "message": "[PATCH] remove duplicate get_dentry functions in various places\n\nVarious filesystem drivers have grown a get_dentry() function that\u0027s a\nduplicate of lookup_one_len, except that it doesn\u0027t take a maximum length\nargument and doesn\u0027t check for \\0 or / in the passed in filename.\n\nSwitch all these places to use lookup_one_len.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nCc: Greg KH \u003cgreg@kroah.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "8215534ce7d073423bfa9c17405c43ab7636ca03", "tree": "b53aed1111cf10a7d42ef0695308c4a70b820747", "parents": [ "988d186de5b6966a71a8cc52e6cb4895fd2f7799" ], "author": { "name": "Maneesh Soni", "email": "maneesh@in.ibm.com", "time": "Tue May 31 10:39:52 2005 +0530" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Mon Jun 20 15:15:37 2005 -0700" }, "message": "[PATCH] sysfs-iattr: set inode attributes\n\no Following patch sets the attributes for newly allocated inodes for sysfs\n objects. If the object has non-default attributes, inode attributes are\n set as saved in sysfs_dirent-\u003es_iattr, pointer to struct iattr.\n\nSigned-off-by: Maneesh Soni \u003cmaneesh@in.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n" }, { "commit": "988d186de5b6966a71a8cc52e6cb4895fd2f7799", "tree": "428cb6c29cbe2563eb91f9f2a03512b7eafa9449", "parents": [ "6fa5c828c7fb6beef7035864bd2b18e7386fbdd5" ], "author": { "name": "Maneesh Soni", "email": "maneesh@in.ibm.com", "time": "Tue May 31 10:39:14 2005 +0530" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Mon Jun 20 15:15:37 2005 -0700" }, "message": "[PATCH] sysfs-iattr: add sysfs_setattr\n\no This adds -\u003ei_op-\u003esetattr VFS method for sysfs inodes. The changed\n attribues are saved in the persistent sysfs_dirent structure as a pointer\n to struct iattr. The struct iattr is allocated only for those sysfs_dirent\u0027s\n for which default attributes are getting changed. Thanks to Jon Smirl for\n this suggestion.\n\nSigned-off-by: Maneesh Soni \u003cmaneesh@in.ibm.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n" }, { "commit": "1da177e4c3f41524e886b7f1b8a0c1fc7321cac2", "tree": "0bba044c4ce775e45a88a51686b5d9f90697ea9d", "parents": [], "author": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "message": "Linux-2.6.12-rc2\n\nInitial git repository build. I\u0027m not bothering with the full history,\neven though we have it. We can create a separate \"historical\" git\narchive of that later if we want to, and in the meantime it\u0027s about\n3.2GB when imported into git - space that would just make the early\ngit days unnecessarily complicated, when we don\u0027t have a lot of good\ninfrastructure for it.\n\nLet it rip!\n" } ] }