)]}' { "log": [ { "commit": "070dc6dd7103b6b3f7e4d46e754354a5c15f366e", "tree": "173e72d08b2070d496c203a6d5156b9a2ad2808c", "parents": [ "bd52b746262c8d77e73903d6608014fb2fcdcd9d" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Aug 30 17:33:34 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Aug 30 18:06:21 2010 +1000" }, "message": "md: resolve confusion of MD_CHANGE_CLEAN\n\nMD_CHANGE_CLEAN is used for two different purposes and this leads to\nconfusion.\nOne of the purposes is largely mirrored by MD_CHANGE_PENDING which is\nnot used for anything else, so have MD_CHANGE_PENDING take over that\npurpose fully.\n\nThe two purposes are:\n 1/ tell md_update_sb that an update is needed and that it is just a\n clean/dirty transition.\n 2/ tell user-space that an transition from clean to dirty is pending\n (something wants to write), and tell te kernel (by clearin the\n flag) that the transition is OK.\n\nThe first purpose remains wit MD_CHANGE_CLEAN, the second is moved\nfully to MD_CHANGE_PENDING.\n\nThis means that various places which conditionally set or cleared\nMD_CHANGE_CLEAN no longer need to be conditional.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "3d30701b58970425e1d45994d6cb82f828924fdd", "tree": "8b14cf462628bebf8548c1b8c205a674564052d1", "parents": [ "8cbd84f2dd4e52a8771b191030c374ba3e56d291", "fd8aa2c1811bf60ccb2d5de0579c6f62aec1772d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 10 15:38:19 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Aug 10 15:38:19 2010 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://neil.brown.name/md\n\n* \u0027for-linus\u0027 of git://neil.brown.name/md: (24 commits)\n md: clean up do_md_stop\n md: fix another deadlock with removing sysfs attributes.\n md: move revalidate_disk() back outside open_mutex\n md/raid10: fix deadlock with unaligned read during resync\n md/bitmap: separate out loading a bitmap from initialising the structures.\n md/bitmap: prepare for storing write-intent-bitmap via dm-dirty-log.\n md/bitmap: optimise scanning of empty bitmaps.\n md/bitmap: clean up plugging calls.\n md/bitmap: reduce dependence on sysfs.\n md/bitmap: white space clean up and similar.\n md/raid5: export raid5 unplugging interface.\n md/plug: optionally use plugger to unplug an array during resync/recovery.\n md/raid5: add simple plugging infrastructure.\n md/raid5: export is_congested test\n raid5: Don\u0027t set read-ahead when there is no queue\n md: add support for raising dm events.\n md: export various start/stop interfaces\n md: split out md_rdev_init\n md: be more careful setting MD_CHANGE_CLEAN\n md/raid5: ensure we create a unique name for kmem_cache when mddev has no gendisk\n ...\n" }, { "commit": "bb4f1e9d0e2ef93de8e36ca0f5f26625fcd70b7d", "tree": "7c7edc0d5fa2b5702358f11396d52d07183708c0", "parents": [ "147e0b6a639ac581ca3bf627bedc3f4a6d3eca66" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Sun Aug 08 21:18:03 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Sun Aug 08 21:21:27 2010 +1000" }, "message": "md: fix another deadlock with removing sysfs attributes.\n\nMove the deletion of sysfs attributes from reconfig_mutex to\nopen_mutex didn\u0027t really help as a process can try to take\nopen_mutex while holding reconfig_mutex, so the same deadlock can\nhappen, just requiring one more process to be involved in the chain.\n\nI looks like I cannot easily use locking to wait for the sysfs\ndeletion to complete, so don\u0027t.\n\nThe only things that we cannot do while the deletions are still\npending is other things which can change the sysfs namespace: run,\ntakeover, stop. Each of these can fail with -EBUSY.\nSo set a flag while doing a sysfs deletion, and fail run, takeover,\nstop if that flag is set.\n\nThis is suitable for 2.6.35.x\n\nCc: stable@kernel.org\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "7b6d91daee5cac6402186ff224c3af39d79f4a0e", "tree": "b1518cf0b6c301178e0a320f80610cd5b3aa7625", "parents": [ "33659ebbae262228eef4e0fe990f393d1f0ed941" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Sat Aug 07 18:20:39 2010 +0200" }, "committer": { "name": "Jens Axboe", "email": "jaxboe@fusionio.com", "time": "Sat Aug 07 18:20:39 2010 +0200" }, "message": "block: unify flags for struct bio and struct request\n\nRemove the current bio flags and reuse the request flags for the bio, too.\nThis allows to more easily trace the type of I/O from the filesystem\ndown to the block driver. There were two flags in the bio that were\nmissing in the requests: BIO_RW_UNPLUG and BIO_RW_AHEAD. Also I\u0027ve\nrenamed two request flags that had a superflous RW in them.\n\nNote that the flags are in bio.h despite having the REQ_ name - as\nblkdev.h includes bio.h that is the only way to go for now.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Jens Axboe \u003cjaxboe@fusionio.com\u003e\n" }, { "commit": "e384e58549a2e9a83071ad80280c1a9053cfd84c", "tree": "805619143612ddc8cc71a45dd4522393c23a576f", "parents": [ "ef4256733506f2459a0c436b62267d22a3f0cec6" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:34 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 13:21:34 2010 +1000" }, "message": "md/bitmap: prepare for storing write-intent-bitmap via dm-dirty-log.\n\nThis allows md/raid5 to fully work as a dm target.\n\nNormally md uses a \u0027filemap\u0027 which contains a list of pages of bits\neach of which may be written separately.\ndm-log uses and all-or-nothing approach to writing the log, so\nwhen using a dm-log, -\u003efilemap is NULL and the flags normally stored\nin filemap_attr are stored in -\u003elogattrs instead.\n\n\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "b63d7c2e29bf9cc94989806f2df0cfca4976b830", "tree": "fdb82527ca458769d3c7381375873931bfb23b19", "parents": [ "5ff5afffe6527543866a47ffab12769427283917" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:33 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 13:21:32 2010 +1000" }, "message": "md/bitmap: clean up plugging calls.\n\n1/ use md_unplug in bitmap.c as we will soon be using bitmaps under\n arrays with no queue attached.\n\n2/ Don\u0027t bother plugging the queue when we set a bit in the bitmap.\n The reason for this was to encourage as many bits as possible to\n get set before we unplug and write stuff out.\n However every personality already plugs the queue after\n bitmap_startwrite either directly (raid1/raid10) or be setting\n STRIPE_BIT_DELAY which causes the queue to be plugged later\n (raid5).\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "ac2f40be46ce6ab3bec4c8c297d6923f941741ce", "tree": "f1abc7d627edaaf450dbff3e68a902ff667a3c92", "parents": [ "9f7c2220017771253d7d10b3cc017cb79eeac0fb" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:31 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 13:07:22 2010 +1000" }, "message": "md/bitmap: white space clean up and similar.\n\nFixes some whitespace problems\nFixed some checkpatch.pl complaints.\nReplaced kmalloc ... memset(0), with kzalloc\nFixed an unlikely memory leak on an error path.\nReformatted a number of \u0027if/else\u0027 sets, sometimes\nreplacing goto with an else clause.\nRemoved some old comments and commented-out code.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "252ac5221a71be72b7e7c7b7482af91e9c962e8c", "tree": "cb9542b1391dabc136ce0052eb28948e77a6e2e5", "parents": [ "2ac8740151b082f045e58010eb92560c3a23a0e9" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:29 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 12:53:08 2010 +1000" }, "message": "md/plug: optionally use plugger to unplug an array during resync/recovery.\n\nIf an array doesn\u0027t have a \u0027queue\u0027 then md_do_sync cannot\nunplug it.\nIn that case it will have a \u0027plugger\u0027, so make that available\nto the mddev, and use it to unplug the array if needed.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "2ac8740151b082f045e58010eb92560c3a23a0e9", "tree": "39bff686e28f033339c6d1cf47042b6dc1586c2f", "parents": [ "11d8a6e3719519fbc0e2c9d61b6fa931b84bf813" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:29 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 12:53:08 2010 +1000" }, "message": "md/raid5: add simple plugging infrastructure.\n\nmd/raid5 uses the plugging infrastructure provided by the block layer\nand \u0027struct request_queue\u0027. However when we plug raid5 under dm there\nis no request queue so we cannot use that.\n\nSo create a similar infrastructure that is much lighter weight and use\nit for raid5.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "768a418db102bb6aa6064e6090892b5c21ff1f9e", "tree": "2f33acbf0ca542f43c8014201b539ae8f461ee5c", "parents": [ "390ee602a142a93f2c7eb7bffee8e277058b8e0a" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 11:49:55 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 12:52:27 2010 +1000" }, "message": "md: add support for raising dm events.\n\ndm uses scheduled work to raise events to user-space.\nSo allow md device to have work_structs and schedule them on an error.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "390ee602a142a93f2c7eb7bffee8e277058b8e0a", "tree": "33a23748dd5ff4ea76b34ed3e458c6da91185376", "parents": [ "e8bb9a839a26f076379e9cb9f46a879d210156f1" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:27 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 12:52:27 2010 +1000" }, "message": "md: export various start/stop interfaces\n\nexport entry points for starting and stopping md arrays.\nThis will be used by a module to make md/raid5 work under\ndm.\nAlso stop calling md_stop_writes from md_stop, as that won\u0027t\nwork well with dm - it will want to call the two separately.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "e8bb9a839a26f076379e9cb9f46a879d210156f1", "tree": "4da184e861cc5fea4967089711dfdc4fbb175f43", "parents": [ "676e42d896ab6967859fabbb06f4e11b9615cbcf" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:26 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 26 12:52:27 2010 +1000" }, "message": "md: split out md_rdev_init\n\nThis functionality will be needed separately in a subsequent patch, so\nsplit it into it\u0027s own exported function.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "00bcb4ac7ee7e557a491b614219142cea0ef16f4", "tree": "992cfa1d0e653aa3de8f0865c5e1a5a0fee4f935", "parents": [ "d0c6f6258478e1dba532bf7c28e2cd6e1047d3a4" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 01 19:37:23 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Jul 21 13:27:53 2010 +1000" }, "message": "md: reduce dependence on sysfs.\n\nWe will want md devices to live as dm targets where sysfs is not\nvisible. So allow md to not connect to sysfs.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "e93f68a1fc6244c05ad8fae28e75835ec74ab34e", "tree": "d282978aac8f6fcec512be2a6e61287bbb6241b1", "parents": [ "0544a21db02c1d8883158fd6f323364f830a120a" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jun 15 09:36:03 2010 +0100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 24 13:33:24 2010 +1000" }, "message": "md: fix handling of array level takeover that re-arranges devices.\n\nMost array level changes leave the list of devices largely unchanged,\npossibly causing one at the end to become redundant.\nHowever conversions between RAID0 and RAID10 need to renumber\nall devices (except 0).\n\nThis renumbering is currently being done in the -\u003erun method when the\nnew personality takes over. However this is too late as the common\ncode in md.c might already have invalidated some of the devices if\nthey had a -\u003eraid_disk number that appeared to high.\n\nMoving it into the -\u003etakeover method is too early as the array is\nstill active at that time and wrong -\u003eraid_disk numbers could cause\nconfusion.\n\nSo add a -\u003enew_raid_disk field to mdk_rdev_s and use it to communicate\nthe new raid_disk number.\nNow the common code knows exactly which devices need to be renumbered,\nand which can be invalidated, and can do it all at a convenient time\nwhen the array is suspend.\nIt can also update some symlinks in sysfs which previously were not be\nupdated correctly.\n\nReported-by: Maciej Trela \u003cmaciej.trela@intel.com\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "a8707c08f4f718bb0ed65499d3f43201f6e41455", "tree": "8fc3ea9ea924ebc4f28012c11f3c1e70ad60bcb2", "parents": [ "7b0bb5368a7195606eca475d9f4e291ab7227052" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue May 18 09:28:43 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue May 18 15:28:01 2010 +1000" }, "message": "md: simplify updating of event count to sometimes avoid updating spares.\n\nWhen updating the event count for a simple clean \u003c-\u003e dirty transition,\nwe try to avoid updating the spares so they can safely spin-down.\nAs the event_counts across an array must be +/- 1, this means\ndecrementing the event_count on a dirty-\u003eclean transition.\nThis is not always safe and we have to avoid the unsafe time.\nWe current do this with a misguided idea about it being safe or\nnot depending on whether the event_count is odd or even. This\napproach only works reliably in a few common instances, but easily\nfalls down.\n\nSo instead, simply keep internal state concerning whether it is safe\nor not, and always assume it is not safe when an array is first\nassembled.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "21a52c6d05c15f862797736393915bfa8cd40ee9", "tree": "44ffb633d209457979177a52132761a368fcb976", "parents": [ "cca9cf90c504d98644ace52c474770970729f0eb" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Apr 01 15:02:13 2010 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue May 18 15:27:55 2010 +1000" }, "message": "md: pass mddev to make_request functions rather than request_queue\n\nWe used to pass the personality make_request function direct\nto the block layer so the first argument had to be a queue.\nBut now we have the intermediary md_make_request so it makes\nat lot more sense to pass a struct mddev_s.\nIt makes it possible to have an mddev without its own queue too.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "b821eaa572fd737faaf6928ba046e571526c36c6", "tree": "0db462a3c7da3dbf42d28e2483a27d78839b4589", "parents": [ "49ce6cea85fb8d25ee59486c919406e9cecf1762" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 29 11:18:15 2010 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue May 18 15:27:53 2010 +1000" }, "message": "md: remove -\u003echanged and related code.\n\nWe set -\u003echanged to 1 and call check_disk_change at the end\nof md_open so that bd_invalidated would be set and thus\npartition rescan would happen appropriately.\n\nNow that we call revalidate_disk directly, which sets bd_invalidates,\nthat indirection is no longer needed and can be removed.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "c0cc75f84e0e413bce2dcabea74ef418da45c7c1", "tree": "f5723fbd0f5828cadaeec650abb8066ad52fad71", "parents": [ "7b92813c3c0b6990f14838e3985fb385d2655d0c" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 22 10:28:51 2010 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue May 18 15:27:47 2010 +1000" }, "message": "md: discard StateChanged device flag.\n\nThis was needed when sysfs files could only be \u0027notified\u0027\nfrom process context. Now that we have sys_notify_direct,\nwe can call it directly from an interrupt.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "ee8b81b03dffa1c0075553d01c557714aedb85a1", "tree": "95ba42aad6fb632699cd9e60d700d05ff96bc889", "parents": [ "964147d5c86d63be79b442c30f3783d49860c078" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 08 16:02:36 2010 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue May 18 15:27:45 2010 +1000" }, "message": "md: remove some dead fields from mddev_s\n\nThese fields have never been used.\ncommit 4b6d287f627b5fb6a49f78f9e81649ff98c62bb7\nadded them, but also added identical files to bitmap_super_s,\nand only used the latter.\n\nSo remove these unused fields.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "a64c876fd357906a1f7193723866562ad290654c", "tree": "077fcdab730d40227585025bf15bf8e027949a29", "parents": [ "b6eb127d274385d81ce8dd45c98190f097bce1b4" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Apr 14 17:15:37 2010 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon May 17 14:45:40 2010 +1000" }, "message": "md: manage redundancy group in sysfs when changing level.\n\nSome levels expect the \u0027redundancy group\u0027 to be present,\nothers don\u0027t.\nSo when we change level of an array we might need to\nadd or remove this group.\n\nThis requires fixing up the current practice of overloading -\u003eprivate\nto indicate (when -\u003epers \u003d\u003d NULL) that something needs to be removed.\nSo create a new -\u003eto_remove to fill that role.\n\nWhen changing levels, we may need to add or remove attributes. When\nchanging RAID5 -\u003e RAID6, we both add and remove the same thing. It is\nimportant to catch this and optimise it out as the removal is delayed\nuntil a lock is released, so trying to add immediately would cause\nproblems.\n\n\nCc: stable@kernel.org\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "1e50915fe0bbf7a46db0fa7e1e604d3fc95f057d", "tree": "7a722ad6f56c61a6173493f1cd44d809c8b1bd8d", "parents": [ "67b8dc4b06b0e97df55fd76e209f34f9a52e820e" ], "author": { "name": "Robert Becker", "email": "Rob.Becker@riverbed.com", "time": "Mon Dec 14 12:49:58 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "raid: improve MD/raid10 handling of correctable read errors.\n\nWe\u0027ve noticed severe lasting performance degradation of our raid\narrays when we have drives that yield large amounts of media errors.\nThe raid10 module will queue each failed read for retry, and also\nwill attempt call fix_read_error() to perform the read recovery.\nRead recovery is performed while the array is frozen, so repeated\nrecovery attempts can degrade the performance of the array for\nextended periods of time.\n\nWith this patch I propose adding a per md device max number of\ncorrected read attempts. Each rdev will maintain a count of\nread correction attempts in the rdev-\u003eread_errors field (not\nused currently for raid10). When we enter fix_read_error()\nwe\u0027ll check to see when the last read error occurred, and\ndivide the read error count by 2 for every hour since the\nlast read error. If at that point our read error count\nexceeds the read error threshold, we\u0027ll fail the raid device.\n\nIn addition in this patch I add sysfs nodes (get/set) for\nthe per md max_read_errors attribute, the rdev-\u003eread_errors\nattribute, and added some printk\u0027s to indicate when\nfix_read_error fails to repair an rdev.\n\nFor testing I used debugfs-\u003efail_make_request to inject\nIO errors to the rdev while doing IO to the raid array.\n\nSigned-off-by: Robert Becker \u003cRob.Becker@riverbed.com\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "ece5cff0da9e696c360fff592cb5f51b6419e4d6", "tree": "1c5cab0e89ae466be9edb476ceb91e5c898fafa6", "parents": [ "624ce4f5658fa3e0303c1217bba2706142fe7568" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:56 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "md: Support write-intent bitmaps with externally managed metadata.\n\nIn this case, the metadata needs to not be in the same\nsector as the bitmap.\nmd will not read/write any bitmap metadata. Config must be\ndone via sysfs and when a recovery makes the array non-degraded\nagain, writing \u0027true\u0027 to \u0027bitmap/can_clear\u0027 will allow bits in\nthe bitmap to be cleared again.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "43a705076e51c5af21ec4260a35699775ea298f5", "tree": "ff8e161d841c41992dbbcca2a844816411941c3c", "parents": [ "72e02075a33f739e21430262f71da8e82db9dbb3" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:55 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "md: support updating bitmap parameters via sysfs.\n\nA new attribute directory \u0027bitmap\u0027 in \u0027md\u0027 is created which\ncontains files for configuring the bitmap.\n\u0027location\u0027 identifies where the bitmap is, either \u0027none\u0027,\nor \u0027file\u0027 or \u0027sector offset from metadata\u0027.\nWriting \u0027location\u0027 can create or remove a bitmap.\nAdding a \u0027file\u0027 bitmap this way is not yet supported.\n\u0027chunksize\u0027 and \u0027time_base\u0027 must be set before \u0027location\u0027\ncan be set.\n\n\u0027chunksize\u0027 can be set before creating a bitmap, but is\ncurrently always over-ridden by the bitmap superblock.\n\n\u0027time_base\u0027 and \u0027backlog\u0027 can be updated at any time.\n\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\nReviewed-by: Andre Noll \u003cmaan@systemlinux.org\u003e\n" }, { "commit": "72e02075a33f739e21430262f71da8e82db9dbb3", "tree": "04a87a16751b714fe0fb41f28ccc54af9f820de1", "parents": [ "f6af949c5672115313cc3c976d85b0533f607d7e" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:55 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "md: factor out parsing of fixed-point numbers\n\nsafe_delay_store can parse fixed point numbers (for fractions\nof a second). We will want to do that for another sysfs\nfile soon, so factor out the code.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "f6af949c5672115313cc3c976d85b0533f607d7e", "tree": "6d00471bd4ee7c3aee67f9ad4c0a3023109b88e0", "parents": [ "9cd30fdc33cde9ae4ac55a1ccbbb89f3f7b9b2f2" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:54 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "md: support bitmap offset appropriate for external-metadata arrays.\n\nFor md arrays were metadata is managed externally, the kernel does not\nknow about a superblock so the superblock offset is 0.\nIf we want to have a write-intent-bitmap near the end of the\ndevices of such an array, we should support sector_t sized offset.\nWe need offset be possibly negative for when the bitmap is before\nthe metadata, so use loff_t instead.\n\nAlso add sanity check that bitmap does not overlap with data.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "42a04b5078ce73a32f85762551d5703c5bd646a1", "tree": "3ef384933cd33d000516c292712da9a99e273360", "parents": [ "c3d9714e88c8685cf9bc837c3241fc005f95fb82" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:53 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "md: move offset, daemon_sleep and chunksize out of bitmap structure\n\n... and into bitmap_info. These are all configuration parameters\nthat need to be set before the bitmap is created.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "c3d9714e88c8685cf9bc837c3241fc005f95fb82", "tree": "dd11e16ec911983bb7452bb4fb092769ba612e3b", "parents": [ "709ae4879ae33628ded276ce7da8cd5acfec476b" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:52 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:51:41 2009 +1100" }, "message": "md: collect bitmap-specific fields into one structure.\n\nIn preparation for making bitmap fields configurable via sysfs,\nstart tidying up by making a single structure to contain the\nconfiguration fields.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "a2826aa92e2e14db372eda01d333267258944033", "tree": "9cdd3329205bf480a4782705a3db1738e3faae44", "parents": [ "efa593390e70b0e3c39f6b2dca8876b6b1461e41" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:49 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:49 2009 +1100" }, "message": "md: support barrier requests on all personalities.\n\nPreviously barriers were only supported on RAID1. This is because\nother levels requires synchronisation across all devices and so needed\na different approach.\nHere is that approach.\n\nWhen a barrier arrives, we send a zero-length barrier to every active\ndevice. When that completes - and if the original request was not\nempty - we submit the barrier request itself (with the barrier flag\ncleared) and then submit a fresh load of zero length barriers.\n\nThe barrier request itself is asynchronous, but any subsequent\nrequest will block until the barrier completes.\n\nThe reason for clearing the barrier flag is that a barrier request is\nallowed to fail. If we pass a non-empty barrier through a striping\nraid level it is conceivable that part of it could succeed and part\ncould fail. That would be way too hard to deal with.\nSo if the first run of zero length barriers succeed, we assume all is\nsufficiently well that we send the request and ignore errors in the\nsecond run of barriers.\n\nRAID5 needs extra care as write requests may not have been submitted\nto the underlying devices yet. So we flush the stripe cache before\nproceeding with the barrier.\n\nNote that the second set of zero-length barriers are submitted\nimmediately after the original request is submitted. Thus when\na personality finds mddev-\u003ebarrier to be set during make_request,\nit should not return from make_request until the corresponding\nper-device request(s) have been queued.\n\nThat will be done in later patches.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\nReviewed-by: Andre Noll \u003cmaan@systemlinux.org\u003e\n" }, { "commit": "aa5cbd103887011b4830355f88fb055f9ad2d556", "tree": "92ef7c80999c6d0bfc6fef9a9cd3e4a75ee16c06", "parents": [ "f40542532e96dda5506eb76badea322f2ae4731c" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:46 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Dec 14 12:49:46 2009 +1100" }, "message": "md/bitmap: protect against bitmap removal while being updated.\n\nA write intent bitmap can be removed from an array while the\narray is active.\nWhen this happens, all IO is suspended and flushed before the\nbitmap is removed.\nHowever it is possible that bitmap_daemon_work is still running to\nclear old bits from the bitmap. If it is, it can dereference the\nbitmap after it has been freed.\n\nSo introduce a new mutex to protect bitmap_daemon_work and get it\nbefore destroying a bitmap.\n\nThis is suitable for any current -stable kernel.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\nCc: stable@kernel.org\n" }, { "commit": "3fa841d7e7266f6fcc1b3885b905f5153ba897d8", "tree": "7afa971a10eddcf52787340ea43fa1e910b5805b", "parents": [ "0da3c6194ec2f32617b272df4505a1cf022faea5" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Sep 23 18:10:29 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Sep 23 18:10:29 2009 +1000" }, "message": "md: report device as congested when suspended\n\nThis should writeback from coming when the device is temporarily\nsuspended.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "411c94038594b2a3fd123d09bdec3fe2500e383d", "tree": "44b991786c7f4806b1b88f40439945b3835b6292", "parents": [ "786d8ca341a30296264bc6cebac52d37b0851647" ], "author": { "name": "Anand Gadiyar", "email": "gadiyar@ti.com", "time": "Tue Jul 07 15:24:23 2009 +0530" }, "committer": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Mon Sep 21 15:14:54 2009 +0200" }, "message": "trivial: fix typo \"for for\" in multiple files\n\ntrivial: fix typo \"for for\" in multiple files\n\nSigned-off-by: Anand Gadiyar \u003cgadiyar@ti.com\u003e\nSigned-off-by: Jiri Kosina \u003cjkosina@suse.cz\u003e\n" }, { "commit": "c8c00a6915a2e3d10416e8bdd3138429beb96210", "tree": "ecf06a9e2b08edefe707da450b52a69f818ec7d6", "parents": [ "7b2aa037e878c939676675969983284a02958ae3" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Aug 10 12:50:52 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Aug 10 12:50:52 2009 +1000" }, "message": "Remove deadlock potential in md_open\n\nA recent commit:\n commit 449aad3e25358812c43afc60918c5ad3819488e7\n\nintroduced the possibility of an A-B/B-A deadlock between\nbd_mutex and reconfig_mutex.\n\n__blkdev_get holds bd_mutex while calling md_open which takes\n reconfig_mutex,\ndo_md_run is always called with reconfig_mutex held, and it now\n takes bd_mutex in the call the revalidate_disk.\n\nThis potential deadlock was not caught by lockdep due to the\nuse of mutex_lock_interruptible_nexted which was introduced\nby\n commit d63a5a74dee87883fda6b7d170244acaac5b05e8\ndo avoid a warning of an impossible deadlock.\n\nIt is quite possible to split reconfig_mutex in to two locks.\nOne protects the array data structures while it is being\nreconfigured, the other ensures that an array is never even partially\nopen while it is being deactivated.\nIn particular, the second lock prevents an open from completing\nbetween the time when do_md_stop checks if there are any active opens,\nand the time when the array is either set read-only, or when -\u003epers is\nset to NULL. So we can be certain that no IO is in flight as the\narray is being destroyed.\n\nSo create a new lock, open_mutex, just to ensure exclusion between\n\u0027open\u0027 and \u0027stop\u0027.\n\nThis avoids the deadlock and also avoids the lockdep warning mentioned\nin commit d63a5a74d\n\nReported-by: \"Mike Snitzer\" \u003csnitzer@gmail.com\u003e\nReported-by: \"H. Peter Anvin\" \u003chpa@zytor.com\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "ac5e7113e74872928844d00085bd47c988f12728", "tree": "dabce6a99c2af9f00b33660f6b2d906cce836764", "parents": [ "95fc17aac45300f45968aacd97a536ddd8db8101" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Mon Aug 03 10:59:47 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Aug 03 10:59:47 2009 +1000" }, "message": "md: Push down data integrity code to personalities.\n\nThis patch replaces md_integrity_check() by two new public functions:\nmd_integrity_register() and md_integrity_add_rdev() which are both\npersonality-independent.\n\nmd_integrity_register() is called from the -\u003erun and -\u003ehot_remove\nmethods of all personalities that support data integrity. The\nfunction iterates over the component devices of the array and\ndetermines if all active devices are integrity capable and if their\nprofiles match. If this is the case, the common profile is registered\nfor the mddev via blk_integrity_register().\n\nThe second new function, md_integrity_add_rdev() is called from the\n-\u003ehot_add_disk methods, i.e. whenever a new device is being added\nto a raid array. If the new device does not support data integrity,\nor has a profile different from the one already registered, data\nintegrity for the mddev is disabled.\n\nFor raid0 and linear, only the call to md_integrity_register() from\nthe -\u003erun method is necessary.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "0894cc3066aaa3e75a99383c0d25feebf9b688ac", "tree": "651664e6b288e36dc9553443e334e650139851c8", "parents": [ "8190e754e0723de7cecb80bdd9eb93911dfa04a1" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Thu Jun 18 08:49:23 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:49:23 2009 +1000" }, "message": "md: Move check for bitmap presence to personality code.\n\nIf the superblock of a component device indicates the presence of a\nbitmap but the corresponding raid personality does not support bitmaps\n(raid0, linear, multipath, faulty), then something is seriously wrong\nand we\u0027d better refuse to run such an array.\n\nCurrently, this check is performed while the superblocks are examined,\ni.e. before entering personality code. Therefore the generic md layer\nmust know which raid levels support bitmaps and which do not.\n\nThis patch avoids this layer violation without adding identical code\nto various personalities. This is accomplished by introducing a new\npublic function to md.c, md_check_no_bitmap(), which replaces the\nhard-coded checks in the superblock loading functions.\n\nA call to md_check_no_bitmap() is added to the -\u003erun method of each\npersonality which does not support bitmaps and assembly is aborted\nif at least one component device contains a bitmap.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "8190e754e0723de7cecb80bdd9eb93911dfa04a1", "tree": "6fbd0edd64cb767b9421103301e198804c3aa40b", "parents": [ "13f2682b7216ebebd72b3d5868fe7fccec91a92d" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:48:58 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:48:58 2009 +1000" }, "message": "md: remove chunksize rounding from common code.\n\nIt is easiest to round sizes to multiples of chunk size in\nthe personality code for those personalities which care.\nThose personalities now do the rounding, so we can\nremove that function from common code.\n\nAlso remove the upper bound on the size of a chunk, and the lower\nbound on the size of a device (1 chunk), neither of which really buy\nus anything.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "50ac168a6e0a061bf5346d53aa9e7beb94c97527", "tree": "4f1eff67793bd8349a1995c921a63e9512be7818", "parents": [ "597a711b69cfff95c4b8f6069037e7ad3fc71f56" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:47:55 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:47:55 2009 +1000" }, "message": "md: merge reconfig and check_reshape methods.\n\nThe difference between these two methods is artificial.\nBoth check that a pending reshape is valid, and perform any\naspect of it that can be done immediately.\n\u0027reconfig\u0027 handles chunk size and layout.\n\u0027check_reshape\u0027 handles raid_disks.\n\nSo make them just one method.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "597a711b69cfff95c4b8f6069037e7ad3fc71f56", "tree": "8c2cfa4c1223827a560de2f14da426d4822151c4", "parents": [ "01ee22b496c41384eaa6dcae983c86d8bc32fbb8" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:47:42 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:47:42 2009 +1000" }, "message": "md: remove unnecessary arguments from -\u003ereconfig method.\n\nPassing the new layout and chunksize as args is not necessary as\nthe mddev has fields for new_check and new_layout.\n\nThis is preparation for combining the check_reshape and reconfig\nmethods\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "664e7c413f1e90eceb0b2596dd73a0832faec058", "tree": "72ea827a0f2a493766d3ea1fd14909c756aa4496", "parents": [ "9d8f0363623b3da12c43007cf77f5e1a4e8a5964" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Thu Jun 18 08:45:27 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:45:27 2009 +1000" }, "message": "md: Convert mddev-\u003enew_chunk to sectors.\n\nA straight-forward conversion which gets rid of some\nmultiplications/divisions/shifts. The patch also introduces a couple\nof new ones, most of which are due to conf-\u003echunk_size still being\nrepresented in bytes. This will be cleaned up in subsequent patches.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "9d8f0363623b3da12c43007cf77f5e1a4e8a5964", "tree": "0fee53971a397ade209dd36c4f1ed50db6450faf", "parents": [ "fbb704efb784e2c8418e34dc3013af76bdd58101" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Thu Jun 18 08:45:01 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Thu Jun 18 08:45:01 2009 +1000" }, "message": "md: Make mddev-\u003echunk_size sector-based.\n\nThis patch renames the chunk_size field to chunk_sectors with the\nimplied change of semantics. Since\n\n\tis_power_of_2(chunk_size) \u003d is_power_of_2(chunk_sectors \u003c\u003c 9)\n\t\t\t\t \u003d is_power_of_2(chunk_sectors)\n\nthese bits don\u0027t need an adjustment for the shift.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "63fe08177f92ce21929df8af6361491b98ad12cd", "tree": "7d6f5aa9251f3120e0911483463e14e1137ed81c", "parents": [ "b0cbc861a3c05e634520b049b5cc27ad6febb51f" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Tue Apr 14 12:01:53 2009 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Apr 14 12:01:53 2009 +1000" }, "message": "md: tiny md.h cleanups\n\n- update inclusion guard and make sure it covers the whole file\n - remove superflous #ifdef CONFIG_BLOCK\n - make sure all required headers are included so that new users aren\u0027t\n required to include others before\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "cea9c22800773cecb1d41f4a6139f9eb6a95368b", "tree": "d070d22390b86768fefb8ba04c64d74993ed28e9", "parents": [ "7ec0547838976d088dfb9cb0adb073e6e8a15aa3" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 15:15:05 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 15:15:05 2009 +1100" }, "message": "md: add explicit method to signal the end of a reshape.\n\nCurrently raid5 (the only module that supports restriping)\nnotices that the reshape has finished be sync_request being\ngiven a large value, and handles any cleanup them.\n\nThis patch changes it so md_check_recovery calls into an\nexplicit finish_reshape method as well.\n\nThe clean-up from sync_request can do things that need to be\ndone promptly, typically things local to the raid5_conf_t\nstructure.\n\nThe \"finish_reshape\" method is called under the mddev_lock\nso it can do things involving reconfiguring the device.\n\nThis allows us to get rid of md_set_array_sectors_locked, which\nwould have caused a deadlock if you tried to stop and array\nwhile a reshape was happening.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "b522adcde9c4d3fb7b579cfa9160d8bde7744be8", "tree": "4208e3dcec3ebdfa6ad9bc153f76129400532717", "parents": [ "1f403624bde3c678a166984b1e6a727a0ce06f2b" ], "author": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Tue Mar 31 15:00:31 2009 +1100" }, "committer": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Tue Mar 31 15:00:31 2009 +1100" }, "message": "md: \u0027array_size\u0027 sysfs attribute\n\nAllow userspace to set the size of the array according to the following\nsemantics:\n\n1/ size must be \u003c\u003d to the size returned by mddev-\u003epers-\u003esize(mddev, 0, 0)\n a) If size is set before the array is running, do_md_run will fail\n if size is greater than the default size\n b) A reshape attempt that reduces the default size to less than the set\n array size should be blocked\n2/ once userspace sets the size the kernel will not change it\n3/ writing \u0027default\u0027 to this attribute returns control of the size to the\n kernel and reverts to the size reported by the personality\n\nAlso, convert locations that need to know the default size from directly\nreading -\u003earray_sectors to \u003cpers\u003e_size. Resync/reshape operations\nalways follow the default size.\n\nFinally, fixup other locations that read a number of 1k-blocks from\nuserspace to use strict_blocks_to_sectors() which checks for unsigned\nlong long to sector_t overflow and blocks to sectors overflow.\n\nReviewed-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: Dan Williams \u003cdan.j.williams@intel.com\u003e\n" }, { "commit": "1f403624bde3c678a166984b1e6a727a0ce06f2b", "tree": "b66ed37091e825bcde1cd0d333d4ebba86772805", "parents": [ "80c3a6ce4ba4470379b9e6a4d9bcd9d2ee26ae03" ], "author": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Tue Mar 31 14:59:03 2009 +1100" }, "committer": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Tue Mar 31 14:59:03 2009 +1100" }, "message": "md: centralize -\u003earray_sectors modifications\n\nGet personalities out of the business of directly modifying\n-\u003earray_sectors. Lays groundwork to introduce policy on when\n-\u003earray_sectors can be modified.\n\nReviewed-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: Dan Williams \u003cdan.j.williams@intel.com\u003e\n" }, { "commit": "80c3a6ce4ba4470379b9e6a4d9bcd9d2ee26ae03", "tree": "de88622f05a7e681ac7bff7714e8f2d8953b66ac", "parents": [ "93ed05e2a56b5ee06c0100a169433b11c4aa5d0b" ], "author": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Tue Mar 17 18:10:40 2009 -0700" }, "committer": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Tue Mar 31 14:57:49 2009 +1100" }, "message": "md: add \u0027size\u0027 as a personality method\n\nIn preparation for giving userspace control over -\u003earray_sectors we need\nto be able to retrieve the \u0027default\u0027 size, and the \u0027anticipated\u0027 size\nwhen a reshape is requested. For personalities that do not reshape emit\na warning if anything but the default size is requested.\n\nIn the raid5 case we need to update -\u003eprevious_raid_disks to make the\nnew \u0027default\u0027 size available.\n\nReviewed-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: Dan Williams \u003cdan.j.williams@intel.com\u003e\n" }, { "commit": "245f46c2c221ef09c7db892f0e3fc2149be42052", "tree": "be7e6f4b7b65b682c43e59ef17f7ed580bc8a82b", "parents": [ "409c57f3801701dfee27a28103dda4831306cb20" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:39:39 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:39:39 2009 +1100" }, "message": "md: add -\u003etakeover method to support changing the personality managing an array\n\nImplement this for RAID6 to be able to \u0027takeover\u0027 a RAID5 array. The\nnew RAID6 will use a layout which places Q on the last device, and\nthat device will be missing.\nIf there are any available spares, one will immediately have Q\nrecovered onto it.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "409c57f3801701dfee27a28103dda4831306cb20", "tree": "430c8e4ebe879b27250e061dc1a1171b12aaadf0", "parents": [ "e0cf8f045b2023b0b3f919ee93eb94345f648434" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:39:39 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:39:39 2009 +1100" }, "message": "md: enable suspend/resume of md devices.\n\nTo be able to change the \u0027level\u0027 of an md/raid array, we need to\nsuspend the device so that no requests are active - then move some\npointers around etc.\n\nThe code already keeps counts of active requests and the -\u003equiesce\nfunction can be used to wait until those counts hit zero.\nHowever the quiesce function blocks new requests once they are all\nready \u0027inside\u0027 the personality module, and that is too late if we want\nto replace the personality modules.\n\nSo make all md requests come in through a common md_make_request\nfunction that keeps track of how many requests have entered the\nmodules but may not yet be on the internal reference counts.\nAllow md_make_request to be blocked when we want to suspend the\ndevice, and make it possible to wait for all those in-transit requests\nto be added to internal lists so that -\u003equiesce can wait for them.\n\nThere is still a problem that when a request completes, we drop the\nref count inside the personality code so there is a short time between\nwhen the refcount hits zero, and when the personality code is no\nlonger being used.\nThe personality code never blocks (schedule or spinlock) between\ndropping the refcount and exiting the routine, so this should be safe\n(as put_module calls synchronize_sched() before unmapping the module\ncode).\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "dd8ac336c13fd8afdb082ebacb1cddd5cf727889", "tree": "0baa8111d7406774525a1053be461135257d73d2", "parents": [ "58c0fed400603a802968b23ddf78f029c5a84e41" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "message": "md: Represent raid device size in sectors.\n\nThis patch renames the \"size\" field of struct mdk_rdev_s to\n\"sectors\" and changes this field to store sectors instead of\nblocks.\n\nAll users of this field, linear.c, raid0.c and md.c, are fixed up\naccordingly which gets rid of many multiplications and divisions.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "58c0fed400603a802968b23ddf78f029c5a84e41", "tree": "474fcb9775bb07f39ebb7802fb9b51d69222dcbb", "parents": [ "575a80fa4f623141e9791e41879d87800fb6d862" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "message": "md: Make mddev-\u003esize sector-based.\n\nThis patch renames the \"size\" field of struct mddev_s to \"dev_sectors\"\nand stores the number of 512-byte sectors instead of the number of\n1K-blocks in it.\n\nAll users of that field, including raid levels 1,4-6,10, are adjusted\naccordingly. This simplifies the code a bit because it allows to get\nrid of a couple of divisions/multiplications by two.\n\nIn order to make checkpatch happy, some minor coding style issues\nhave also been addressed. In particular, size_store() now uses\nstrict_strtoull() instead of simple_strtoull().\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "97e4f42d62badb0f9fbc27c013e89bc1336a03bc", "tree": "04cc809702a6b080c417c4ddf605642bbf7de521", "parents": [ "43b2e5d86d8bdd77386226db0bc961529492c043" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "message": "md: occasionally checkpoint drive recovery to reduce duplicate effort after a crash\n\nVersion 1.x metadata has the ability to record the status of a\npartially completed drive recovery.\nHowever we only update that record on a clean shutdown.\nIt would be nice to update it on unclean shutdowns too, particularly\nwhen using a bitmap that removes much to the \u0027sync\u0027 effort after an\nunclean shutdown.\n\nOne complication with checkpointing recovery is that we only know\nwhere we are up to in terms of IO requests started, not which ones\nhave completed. And we need to know what has completed to record\nhow much is recovered. So occasionally pause the recovery until all\nsubmitted requests are completed, then update the record of where\nwe are up to.\n\nWhen we have a bitmap, we already do that pause occasionally to keep\nthe bitmap up-to-date. So enhance that code to record the recovery\noffset and schedule a superblock update.\nAnd when there is no bitmap, just pause 16 times during the resync to\ndo a checkpoint.\n\u002716\u0027 is a fairly arbitrary number. But we don\u0027t really have any good\nway to judge how often is acceptable, and it seems like a reasonable\nnumber for now.\n\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "43b2e5d86d8bdd77386226db0bc961529492c043", "tree": "601a885b1410324678cf319eee77cc54a02f2cf3", "parents": [ "bff61975b3d6c18ee31457cc5b4d73042f44915f" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "message": "md: move md_k.h from include/linux/raid/ to drivers/md/\n\nIt really is nicer to keep related code together..\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "bff61975b3d6c18ee31457cc5b4d73042f44915f", "tree": "3aff48088b35172e74f56ae54f0b53e76a0c2150", "parents": [ "92022950c6b1bb3da90b2976b20271cdfd98b8a3" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "message": "md: move lots of #include lines out of .h files and into .c\n\nThis makes the includes more explicit, and is preparation for moving\nmd_k.h to drivers/md/md.h\n\nRemove include/raid/md.h as its only remaining use was to #include\nother files.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "92022950c6b1bb3da90b2976b20271cdfd98b8a3", "tree": "5405257242ddcadef6e90ea597289558de4e03f2", "parents": [ "8b2b5c217c20b5460218ab8731295f2e46c7dd29" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:33:13 2009 +1100" }, "message": "md: move most content from md.h to md_k.h\n\nThe extern function definitions are kernel-internal definitions, so\nthey belong in md_k.h\n\nThe MD_*_VERSION values could reasonably go in a number of places,\nbut md_u.h seems most reasonable.\n\nThis leaves almost nothing in md.h. It will go soon.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "8b2b5c217c20b5460218ab8731295f2e46c7dd29", "tree": "f5da8b06b3d2a53e80c31e0c0aa386e381b3a841", "parents": [ "ef740c372dfd80e706dbf955d4e4aedda6c0c148" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:27:03 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:27:03 2009 +1100" }, "message": "md: move LEVEL_* definition from md_k.h to md_u.h\n\n.. as they are part of the user-space interface.\nAlso move MdpMinorShift into there so we can remove duplication.\n\nLastly move mdp_major in. It is less obviously part of the user-space\ninterface, but do_mounts_md.c uses it, and it is acting a bit like\nuser-space.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "eea1bf384e05b5ab747f8530c4fba9e9e6907fff", "tree": "ed4c05f8bd7e5805092e9c6a2c257ca71dc0fe2d", "parents": [ "99adcd9d67aaf04e28f5ae96df280f236bde4b66" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:27:02 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 31 14:27:02 2009 +1100" }, "message": "md: Fix is_mddev_idle test (again).\n\nThere are two problems with is_mddev_idle.\n\n1/ sync_io is \u0027atomic_t\u0027 and hence \u0027int\u0027. curr_events and all the\n rest are \u0027long\u0027.\n So if sync_io were to wrap on a 64bit host, the value of\n curr_events would go very negative suddenly, and take a very\n long time to return to positive.\n\n So do all calculations as \u0027int\u0027. That gives us plenty of precision\n for what we need.\n\n2/ To initialise rdev-\u003elast_events we simply call is_mddev_idle, on\n the assumption that it will make sure that last_events is in a\n suitable range. It used to do this, but now it does not.\n So now we need to be more explicit about initialisation.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "4044ba58dd15cb01797c4fd034f39ef4a75f7cc3", "tree": "e1bea8143538fc3eaeeb8578c2f9231e32809a25", "parents": [ "efeb53c0e57213e843b7ef3cc6ebcdea7d6186ac" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:11 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:11 2009 +1100" }, "message": "md: don\u0027t retry recovery of raid1 that fails due to error on source drive.\n\nIf a raid1 has only one working drive and it has a sector which\ngives an error on read, then an attempt to recover onto a spare will\nfail, but as the single remaining drive is not removed from the\narray, the recovery will be immediately re-attempted, resulting\nin an infinite recovery loop.\n\nSo detect this situation and don\u0027t retry recovery once an error\non the lone remaining drive is detected.\n\nAllow recovery to be retried once every time a spare is added\nin case the problem wasn\u0027t actually a media error.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "efeb53c0e57213e843b7ef3cc6ebcdea7d6186ac", "tree": "ef580effd8809fb0bc5d6ffa6f3b8abaa9180539", "parents": [ "d3374825ce57ba2214d375023979f6197ccc1385" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:10 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:10 2009 +1100" }, "message": "md: Allow md devices to be created by name.\n\nUsing sequential numbers to identify md devices is somewhat artificial.\nUsing names can be a lot more user-friendly.\n\nAlso, creating md devices by opening the device special file is a bit\nawkward.\n\nSo this patch provides a new option for creating and naming devices.\n\nWriting a name such as \"md_home\" to\n /sys/modules/md_mod/parameters/new_array\nwill cause an array with that name to be created. It will appear in\n/sys/block/ /proc/partitions and /proc/mdstat as \u0027md_home\u0027.\nIt will have an arbitrary minor number allocated.\n\nmd devices that a created by an open are destroyed on the last\nclose when the device is inactive.\nFor named md devices, they will not be destroyed until the array\nis explicitly stopped, either with the STOP_ARRAY ioctl or by\nwriting \u0027clear\u0027 to /sys/block/md_XXXX/md/array_state.\n\nThe name of the array must start \u0027md_\u0027 to avoid conflict with\nother devices.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "d3374825ce57ba2214d375023979f6197ccc1385", "tree": "441ea927a7c702e4eadeafbac8be97d664bfb83b", "parents": [ "a21d15042d8cd736caf82c2bac564f3f93f3d017" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:10 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:10 2009 +1100" }, "message": "md: make devices disappear when they are no longer needed.\n\nCurrently md devices, once created, never disappear until the module\nis unloaded. This is essentially because the gendisk holds a\nreference to the mddev, and the mddev holds a reference to the\ngendisk, this a circular reference.\n\nIf we drop the reference from mddev to gendisk, then we need to ensure\nthat the mddev is destroyed when the gendisk is destroyed. However it\nis not possible to hook into the gendisk destruction process to enable\nthis.\n\nSo we drop the reference from the gendisk to the mddev and destroy the\ngendisk when the mddev gets destroyed. However this has a\ncomplication.\nBetween the call\n __blkdev_get-\u003eget_gendisk-\u003ekobj_lookup-\u003emd_probe\nand the call\n __blkdev_get-\u003emd_open\n\nthere is no obvious way to hold a reference on the mddev any more, so\nunless something is done, it will disappear and gendisk will be\ndestroyed prematurely.\n\nAlso, once we decide to destroy the mddev, there will be an unlockable\nmoment before the gendisk is unlinked (blk_unregister_region) during\nwhich a new reference to the gendisk can be created. We need to\nensure that this reference can not be used. i.e. the -\u003eopen must\nfail.\n\nSo:\n 1/ in md_probe we set a flag in the mddev (hold_active) which\n indicates that the array should be treated as active, even\n though there are no references, and no appearance of activity.\n This is cleared by md_release when the device is closed if it\n is no longer needed.\n This ensures that the gendisk will survive between md_probe and\n md_open.\n\n 2/ In md_open we check if the mddev we expect to open matches\n the gendisk that we did open.\n If there is a mismatch we return -ERESTARTSYS and modify\n __blkdev_get to retry from the top in that case.\n In the -ERESTARTSYS sys case we make sure to wait until\n the old gendisk (that we succeeded in opening) is really gone so\n we loop at most once.\n\nSome udev configurations will always open an md device when it first\nappears. If we allow an md device that was just created by an open\nto disappear on an immediate close, then this can race with such udev\nconfigurations and result in an infinite loop the device being opened\nand closed, then re-open due to the \u0027ADD\u0027 even from the first open,\nand then close and so on.\nSo we make sure an md device, once created by an open, remains active\nat least until some md \u0027ioctl\u0027 has been made on it. This means that\nall normal usage of md devices will allow them to disappear promptly\nwhen not needed, but the worst that an incorrect usage will do it\ncause an inactive md device to be left in existence (it can easily be\nremoved).\n\nAs an array can be stopped by writing to a sysfs attribute\n echo clear \u003e /sys/block/mdXXX/md/array_state\nwe need to use scheduled work for deleting the gendisk and other\nkobjects. This allows us to wait for any pending gendisk deletion to\ncomplete by simply calling flush_scheduled_work().\n\n\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "159ec1fc060ab22b157a62364045f5e98749c4d3", "tree": "1de0edfd782245b271d2898e36ae76c00e1e1b6d", "parents": [ "ccacc7d2cf03114a24ab903f710118e9e5d43273" ], "author": { "name": "Cheng Renquan", "email": "crquan@gmail.com", "time": "Fri Jan 09 08:31:08 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:08 2009 +1100" }, "message": "md: use list_for_each_entry macro directly\n\nThe rdev_for_each macro defined in \u003clinux/raid/md_k.h\u003e is identical to\nlist_for_each_entry_safe, from \u003clinux/list.h\u003e, it should be defined to\nuse list_for_each_entry_safe, instead of reinventing the wheel.\n\nBut some calls to each_entry_safe don\u0027t really need a safe version,\njust a direct list_for_each_entry is enough, this could save a temp\nvariable (tmp) in every function that used rdev_for_each.\n\nIn this patch, most rdev_for_each loops are replaced by list_for_each_entry,\ntotally save many tmp vars; and only in the other situations that will call\nlist_del to delete an entry, the safe version is used.\n\nSigned-off-by: Cheng Renquan \u003ccrquan@gmail.com\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "0c3573f19d135d718264e38c46597295bd6154b7", "tree": "0943035f89d163d9d7b354445303fb03e5ea7fb1", "parents": [ "538452700d95480c16e7aa6b10ff77cd937d33f4" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:05 2009 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 09 08:31:05 2009 +1100" }, "message": "md: use sysfs_notify_dirent to notify changes to md/sync_action.\n\nThere is no compelling need for this, but sysfs_notify_dirent is a\nnicer interface and the change is good for consistency.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "3c0ee63a64a20351ed6c16ec797e1f8c850741ea", "tree": "482b825a00356a2b7b605065ce1d6ebe87591eb5", "parents": [ "b62b75905d571c29262a6c38cf9e5f089c203871" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Oct 21 13:25:28 2008 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Oct 21 13:25:28 2008 +1100" }, "message": "md: use sysfs_notify_dirent to notify changes to md/dev-xxx/state\n\nThe \u0027state\u0027 file for a device reports, for example, when the device\nhas failed. Changes should be reported to userspace ASAP without\nthe possibility of blocking on low-memory. sysfs_notify does\nhave that possibility (as it takes a mutex which can be held\nacross a kmalloc) so use sysfs_notify_dirent instead.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "b62b75905d571c29262a6c38cf9e5f089c203871", "tree": "0ca9a74c1d53aad0532aaebebb031b9d8edc5d90", "parents": [ "0cfd81031a26717fe14380d18275f8e217571615" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Oct 21 13:25:21 2008 +1100" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Oct 21 13:25:21 2008 +1100" }, "message": "md: use sysfs_notify_dirent to notify changes to md/array_state\n\nNow that we have sysfs_notify_dirent, use it to notify changes\nto md/array_state.\nAs sysfs_notify_dirent can be called in atomic context, we can\nremove the delayed notify and the MD_NOTIFY_ARRAY_STATE flag.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "d8e64406a037a64444175730294e449c9e21f5ec", "tree": "5ac7525899ebc1d2fd9a83a6e19cd0b1b1b1711d", "parents": [ "2339788376e2d69a9154130e4dacd5b21ce63094" ], "author": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Wed Jul 23 13:09:48 2008 -0700" }, "committer": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Wed Jul 23 13:09:48 2008 -0700" }, "message": "md: delay notification of \u0027active_idle\u0027 to the recovery thread\n\nsysfs_notify might sleep, so do not call it from md_safemode_timeout.\n\nSigned-off-by: Dan Williams \u003cdan.j.williams@intel.com\u003e\n" }, { "commit": "4b80991c6cb9efa607bc4fd6f3ecdf5511c31bb0", "tree": "5e2ba7d509af245c29bdf04b00960cc367972c44", "parents": [ "f2ea68cf42aafdd93393b6b8b20fc3c2b5f4390c" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 21 17:05:25 2008 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 21 17:05:25 2008 +1000" }, "message": "md: Protect access to mddev-\u003edisks list using RCU\n\nAll modifications and most access to the mddev-\u003edisks list are made\nunder the reconfig_mutex lock. However there are three places where\nthe list is walked without any locking. If a reconfig happens at this\ntime, havoc (and oops) can ensue.\n\nSo use RCU to protect these accesses:\n - wrap them in rcu_read_{,un}lock()\n - use list_for_each_entry_rcu\n - add to the list with list_add_rcu\n - delete from the list with list_del_rcu\n - delay the \u0027free\u0027 with call_rcu rather than schedule_work\n\nNote that export_rdev did a list_del_init on this list. In almost all\ncases the entry was not in the list anymore so it was a no-op and so\nsafe. It is no longer safe as after list_del_rcu we may not touch\nthe list_head.\nAn audit shows that export_rdev is called:\n - after unbind_rdev_from_array, in which case the delete has\n already been done,\n - after bind_rdev_to_array fails, in which case the delete isn\u0027t needed.\n - before the device has been put on a list at all (e.g. in\n add_new_disk where reading the superblock fails).\n - and in autorun devices after a failure when the device is on a\n different list.\n\nSo remove the list_del_init call from export_rdev, and add it back\nimmediately before the called to export_rdev for that last case.\n\nNote also that -\u003esame_set is sometimes used for lists other than\nmddev-\u003elist (e.g. candidates). In these cases rcu is not needed.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "f2ea68cf42aafdd93393b6b8b20fc3c2b5f4390c", "tree": "00b025b91898ff32dba742b5075dd290fec4dc91", "parents": [ "d6e2215052810678bc9782fd980b52706fc71f50" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 21 17:05:25 2008 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 21 17:05:25 2008 +1000" }, "message": "md: only count actual openers as access which prevent a \u0027stop\u0027\n\nOpen isn\u0027t the only thing that increments -\u003eactive. e.g. reading\n/proc/mdstat will increment it briefly. So to avoid false positives\nin testing for concurrent access, introduce a new counter that counts\njust the number of times the md device it open.\n\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "f233ea5c9e0d8b95e4283bf6a3436b88f6fd3586", "tree": "8a36b621de46da23f3957c6e3ea4d653c2937544", "parents": [ "15f4a5fdf3aa07b53f6a7969664741db5882e485" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Mon Jul 21 17:05:22 2008 +1000" }, "committer": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 21 17:05:22 2008 +1000" }, "message": "md: Make mddev-\u003earray_size sector-based.\n\nThis patch renames the array_size field of struct mddev_s to array_sectors\nand converts all instances to use units of 512 byte sectors instead of 1k\nblocks.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: NeilBrown \u003cneilb@suse.de\u003e\n" }, { "commit": "0f420358e3a2abc028320ace7783e2e38cae77bf", "tree": "99b068212ad1dd8c5e97b151d796d66f53d1f3bb", "parents": [ "b73df2d3d629aefa187a0a3574fd81455e026bc8" ], "author": { "name": "Andre Noll", "email": "maan@systemlinux.org", "time": "Fri Jul 11 22:02:23 2008 +1000" }, "committer": { "name": "Neil Brown", "email": "neilb@suse.de", "time": "Fri Jul 11 22:02:23 2008 +1000" }, "message": "md: Turn rdev-\u003esb_offset into a sector-based quantity.\n\nRename it to sb_start to make sure all users have been converted.\n\nSigned-off-by: Andre Noll \u003cmaan@systemlinux.org\u003e\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\n" }, { "commit": "526647320e696f434647f38421a6ecf65b859c43", "tree": "aa6bf13e6aa766051ba32a8b64157f4adf9fcd3e", "parents": [ "a99ac97113d5bc25ddc4d17f404c2024ac6c57f9" ], "author": { "name": "Neil Brown", "email": "neilb@notabene.brown", "time": "Sat Jun 28 08:31:44 2008 +1000" }, "committer": { "name": "Neil Brown", "email": "neilb@notabene.brown", "time": "Sat Jun 28 08:31:44 2008 +1000" }, "message": "Make sure all changes to md/dev-XX/state are notified\n\nThe important state change happens during an interrupt\nin md_error. So just set a flag there and call sysfs_notify\nlater in process context.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\n" }, { "commit": "72a23c211e4587859d5bf61ac4962d76e593fb02", "tree": "b35b554d7eb9c4b3a2cbc4d9378d362e5e56e44f", "parents": [ "0fd62b861eac7d2dea9b7e939953b20f37186ea1" ], "author": { "name": "Neil Brown", "email": "neilb@notabene.brown", "time": "Sat Jun 28 08:31:41 2008 +1000" }, "committer": { "name": "Neil Brown", "email": "neilb@notabene.brown", "time": "Sat Jun 28 08:31:41 2008 +1000" }, "message": "Make sure all changes to md/sync_action are notified.\n\nWhen the \u0027resync\u0027 thread starts or stops, when we explicitly\nset sync_action, or when we determine that there is definitely nothing\nto do, we notify sync_action.\n\nTo stop \"sync_action\" from occasionally showing the wrong value,\nwe introduce a new flags - MD_RECOVERY_RECOVER - to say that a\nrecovery is probably needed or happening, and we make sure\nthat we set MD_RECOVERY_RUNNING before clearing MD_RECOVERY_NEEDED.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\n" }, { "commit": "5e96ee65c8bd629ce093da67a066d3946468298a", "tree": "e1ff9e4984d71ffaa842e7e1d19c282fa9e01bcd", "parents": [ "a0da84f35b25875870270d16b6eccda4884d61a7" ], "author": { "name": "Neil Brown", "email": "neilb@notabene.brown", "time": "Sat Jun 28 08:31:24 2008 +1000" }, "committer": { "name": "Neil Brown", "email": "neilb@notabene.brown", "time": "Sat Jun 28 08:31:24 2008 +1000" }, "message": "Allow setting start point for requested check/repair\n\nThis makes it possible to just resync a small part of an array.\ne.g. if a drive reports that it has questionable sectors,\na \u0027repair\u0027 of just the region covering those sectors will\ncause them to be read and, if there is an error, re-written\nwith correct data.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\n" }, { "commit": "dfc7064500061677720fa26352963c772d3ebe6b", "tree": "a8ca495bccf98837c6762ffba54a8009c9772259", "parents": [ "90b08710e41a07d4ff0fb8940dcce3a552991a56" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri May 23 13:04:39 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 24 09:56:10 2008 -0700" }, "message": "md: restart recovery cleanly after device failure.\n\nWhen we get any IO error during a recovery (rebuilding a spare), we abort\nthe recovery and restart it.\n\nFor RAID6 (and multi-drive RAID1) it may not be best to restart at the\nbeginning: when multiple failures can be tolerated, the recovery may be\nable to continue and re-doing all that has already been done doesn\u0027t make\nsense.\n\nWe already have the infrastructure to record where a recovery is up to\nand restart from there, but it is not being used properly.\nThis is because:\n - We sometimes abort with MD_RECOVERY_ERR rather than just MD_RECOVERY_INTR,\n which causes the recovery not be be checkpointed.\n - We remove spares and then re-added them which loses important state\n information.\n\nThe distinction between MD_RECOVERY_ERR and MD_RECOVERY_INTR really isn\u0027t\nneeded. If there is an error, the relevant drive will be marked as\nFaulty, and that is enough to ensure correct handling of the error. So we\nfirst remove MD_RECOVERY_ERR, changing some of the uses of it to\nMD_RECOVERY_INTR.\n\nThen we cause the attempt to remove a non-faulty device from an array to\nfail (unless recovery is impossible as the array is too degraded). Then\nwhen remove_and_add_spares attempts to remove the devices on which\nrecovery can continue, it will fail, they will remain in place, and\nrecovery will continue on them as desired.\n\nIssue: If we are halfway through rebuilding a spare and another drive\nfails, and a new spare is immediately available, do we want to:\n 1/ complete the current rebuild, then go back and rebuild the new spare or\n 2/ restart the rebuild from the start and rebuild both devices in\n parallel.\n\nBoth options can be argued for. The code currently takes option 2 as\n a/ this requires least code change\n b/ this results in a minimally-degraded array in minimal time.\n\nCc: \"Eivind Sarto\" \u003civan@kasenna.com\u003e\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "90b08710e41a07d4ff0fb8940dcce3a552991a56", "tree": "8e45d1c6d9b4020099fd36781065bf8b8fdb76e4", "parents": [ "4f54b0e9485644a3c5fca2ae43bcbe7376825747" ], "author": { "name": "Bernd Schubert", "email": "bs@q-leap.de", "time": "Fri May 23 13:04:38 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 24 09:56:10 2008 -0700" }, "message": "md: allow parallel resync of md-devices.\n\nIn some configurations, a raid6 resync can be limited by CPU speed\n(Calculating P and Q and moving data) rather than by device speed. In\nthese cases there is nothing to be gained byt serialising resync of arrays\nthat share a device, and doing the resync in parallel can provide benefit.\n So add a sysfs tunable to flag an array as being allowed to resync in\nparallel with other arrays that use (a different part of) the same device.\n\nSigned-off-by: Bernd Schubert \u003cbs@q-leap.de\u003e\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6bfe0b499082fd3950429017cd8ebf2a6c458aa5", "tree": "81476cf7f7ddbea135bdb93729e0bffae0e7c163", "parents": [ "11e2ede0228ee0f81ccacd15894908c3bf241f73" ], "author": { "name": "Dan Williams", "email": "dan.j.williams@intel.com", "time": "Wed Apr 30 00:52:32 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 30 08:29:33 2008 -0700" }, "message": "md: support blocking writes to an array on device failure\n\nAllows a userspace metadata handler to take action upon detecting a device\nfailure.\n\nBased on an original patch by Neil Brown.\n\nChanges:\n-added blocked_wait waitqueue to rdev\n-don\u0027t qualify Blocked with Faulty always let userspace block writes\n-added md_wait_for_blocked_rdev to wait for the block device to be clear, if\n userspace misses the notification another one is sent every 5 seconds\n-set MD_RECOVERY_NEEDED after clearing \"blocked\"\n-kill DoBlock flag, just test mddev-\u003eexternal\n\nSigned-off-by: Dan Williams \u003cdan.j.williams@intel.com\u003e\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d0fae18f1b53a1d39135a968792be034bdf7ff26", "tree": "3448dab652696848df47dcafb7772220b9e58e4f", "parents": [ "a1801f858e57f87a7f79914346921cc729632295" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Mar 04 14:29:31 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue Mar 04 16:35:18 2008 -0800" }, "message": "md: clean up irregularity with raid autodetect\n\nWhen a raid1 array is stopped, all components currently get added to the list\nfor auto-detection. However we should really only add components that were\nfound by autodetection in the first place. So add a flag to record that\ninformation, and use it.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "73c34431c7119d0bc7d3436abfad75fe47b2c51f", "tree": "c91210130d1115354f011e55c98db9361d4416f8", "parents": [ "d089c6af10c2be5988f03667d6d22fe6085fbe5e" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Feb 06 01:39:59 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Feb 06 10:41:19 2008 -0800" }, "message": "md: change ITERATE_RDEV_GENERIC to rdev_for_each_list, and remove ITERATE_RDEV_PENDING.\n\nFinish ITERATE_ to for_each conversion.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d089c6af10c2be5988f03667d6d22fe6085fbe5e", "tree": "815eb0596bd7c908f30c268da6b5c5eff0e0bc21", "parents": [ "29ac4aa3fc68a86279aca50f20df4d614de2e204" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Feb 06 01:39:59 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Feb 06 10:41:19 2008 -0800" }, "message": "md: change ITERATE_RDEV to rdev_for_each\n\nAs this is more in line with common practice in the kernel. Also swap the\nargs around to be more like list_for_each.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c5d79adba7ced41d7ac097c2ab74759d10522dd5", "tree": "ab5a45046244602f2a27be0b34fb26155a3c8dee", "parents": [ "1ec4a9398dc05061b6258061676fede733458893" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Feb 06 01:39:54 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Feb 06 10:41:18 2008 -0800" }, "message": "md: allow devices to be shared between md arrays\n\nCurrently, a given device is \"claimed\" by a particular array so that it cannot\nbe used by other arrays.\n\nThis is not ideal for DDF and other metadata schemes which have their own\npartitioning concept.\n\nSo for externally managed metadata, just claim the device for md in general,\nrequire that \"offset\" and \"size\" are set properly for each device, and make\nsure that if a device is included in different arrays then the active sections\ndo not overlap.\n\nThis involves adding another flag to the rdev which makes it awkward to set\n\"-\u003eflags \u003d 0\" to clear certain flags. So now clear flags explicitly by name\nwhen we want to clear things.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c620727779f7cc8ea96efb71f0651a26349e59c1", "tree": "777abdad9c9ef10cb4df5c0efc736e6c64851ed8", "parents": [ "c303da6d713b87b7b3f999f5acce8ecc76ff1adb" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Feb 06 01:39:52 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Feb 06 10:41:18 2008 -0800" }, "message": "md: allow a maximum extent to be set for resyncing\n\nThis allows userspace to control resync/reshape progress and synchronise it\nwith other activities, such as shared access in a SAN, or backing up critical\nsections during a tricky reshape.\n\nWriting a number of sectors (which must be a multiple of the chunk size if\nsuch is meaningful) causes a resync to pause when it gets to that point.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "e691063a61f7f72a7d2882eb744b07a520cde23b", "tree": "4f5ceb7ed310a7d1bb076271926260723ac5ee6d", "parents": [ "b47490c9bc73d0b34e4c194db40de183e592e446" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Feb 06 01:39:51 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Feb 06 10:41:18 2008 -0800" }, "message": "md: support \u0027external\u0027 metadata for md arrays\n\n- Add a state flag \u0027external\u0027 to indicate that the metadata is managed\n externally (by user-space) so important changes need to be\n left of user-space to handle.\n Alternates are non-persistant (\u0027none\u0027) where there is no stable metadata -\n after the array is stopped there is no record of it\u0027s status - and\n internal which can be version 0.90 or version 1.x\n These are selected by writing to the \u0027metadata\u0027 attribute.\n\n- move the updating of superblocks (sync_sbs) to after we have checked if\n there are any superblocks or not.\n\n- New array state \u0027write_pending\u0027. This means that the metadata records\n the array as \u0027clean\u0027, but a write has been requested, so the metadata has\n to be updated to record a \u0027dirty\u0027 array before the write can continue.\n This change is reported to md by writing \u0027active\u0027 to the array_state\n attribute.\n\n- tidy up marking of sb_dirty:\n - don\u0027t set sb_dirty when resync finishes as md_check_recovery\n calls md_update_sb when the sync thread finishes anyway.\n - Don\u0027t set sb_dirty in multipath_run as the array might not be dirty.\n - don\u0027t mark superblock dirty when switching to \u0027clean\u0027 if there\n is no internal superblock (if external, userspace can choose to\n update the superblock whenever it chooses to).\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "165125e1e480f9510a5ffcfbfee4e3ee38c05f23", "tree": "8009c8a5ff09e26dc2418d42f66ecafb055c52a2", "parents": [ "f695baf2df9e0413d3521661070103711545207a" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Tue Jul 24 09:28:11 2007 +0200" }, "committer": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Tue Jul 24 09:28:11 2007 +0200" }, "message": "[BLOCK] Get rid of request_queue_t typedef\n\nSome of the code has been gradually transitioned to using the proper\nstruct request_queue, but there\u0027s lots left. So do a full sweet of\nthe kernel and get rid of this typedef and replace its uses with\nthe proper type.\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "713f6ab18b0e7d39f14401362bfe8015b1aedde1", "tree": "fc6ef865d09de1ec816ed96860ffc27dd48d837b", "parents": [ "df968c4e8d809deb2602fb0e8db758f26a4006b9" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Jul 17 04:06:12 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue Jul 17 10:23:15 2007 -0700" }, "message": "md: improve the is_mddev_idle test fix\n\nDon\u0027t use \u0027unsigned\u0027 variable to track sync vs non-sync IO, as the only thing\nwe want to do with them is a signed comparison, and fix up the comment which\nhad become quite wrong.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "44ce6294d07555c3d313757105fd44b78208407f", "tree": "29659bee16a4d7111124c6f41289bc9c11669e70", "parents": [ "497f050c42e46a4b1f6a9bcd8827fa5d97fe1feb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed May 09 18:51:36 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed May 09 18:51:36 2007 -0700" }, "message": "Revert \"md: improve partition detection in md array\"\n\nThis reverts commit 5b479c91da90eef605f851508744bfe8269591a0.\n\nQuoth Neil Brown:\n\n \"It causes an oops when auto-detecting raid arrays, and it doesn\u0027t\n seem easy to fix.\n\n The array may not be \u0027open\u0027 when do_md_run is called, so\n bdev-\u003ebd_disk might be NULL, so bd_set_size can oops.\n\n This whole approach of opening an md device before it has been\n assembled just seems to get more and more painful. I think I\u0027m going\n to have to come up with something clever to provide both backward\n comparability with usage expectation, and sane integration into the\n rest of the kernel.\"\n\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5b479c91da90eef605f851508744bfe8269591a0", "tree": "d7c2844926d96d794d66b974b300eba2d047465e", "parents": [ "08a02ecd28bad35a47357e611044dcbeab06e3d7" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed May 09 02:35:39 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed May 09 12:30:57 2007 -0700" }, "message": "md: improve partition detection in md array\n\nmd currently uses -\u003emedia_changed to make sure rescan_partitions\nis call on md array after they are assembled.\n\nHowever that doesn\u0027t happen until the array is opened, which is later\nthan some people would like.\n\nSo use blkdev_ioctl to do the rescan immediately that the\narray has been assembled.\n\nThis means we can remove all the -\u003echange infrastructure as it was only used\nto trigger a partition rescan.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5792a2856a63cdc568e08a7d6f9b2413d9217b3e", "tree": "042a09fbfdf75dac5efa3bf61c7d6c6dd4c6ba7f", "parents": [ "456a09dce9ca9b0013cabcda918aee851a04471d" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Wed Apr 04 19:08:18 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Apr 04 21:12:47 2007 -0700" }, "message": "[PATCH] md: avoid a deadlock when removing a device from an md array via sysfs\n\nA device can be removed from an md array via e.g.\n echo remove \u003e /sys/block/md3/md/dev-sde/state\n\nThis will try to remove the \u0027dev-sde\u0027 subtree which will deadlock\nsince\n commit e7b0d26a86943370c04d6833c6edba2a72a6e240\n\nWith this patch we run the kobject_del via schedule_work so as to\navoid the deadlock.\n\nCc: Alan Stern \u003cstern@rowland.harvard.edu\u003e\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "e8703fe1f5cdcff686f7eb0a46487b5a04a9324a", "tree": "3b52bfa6ee57d59e059fd0e8d8504dbd3ee073b4", "parents": [ "d33a56d3639eba077489c937943a32ef6237b1b8" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Oct 03 01:15:59 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Tue Oct 03 08:04:18 2006 -0700" }, "message": "[PATCH] md: remove MAX_MD_DEVS which is an arbitrary limit\n\nOnce upon a time we needed to fixed limit to the number of md devices,\nprobably because we preallocated some array. This need no longer exists, but\nwe still have an arbitrary limit.\n\nSo remove MAX_MD_DEVS and allow as many devices as we can fit into the \u0027minor\u0027\npart of a device number.\n\nAlso remove some useless noise at init time (which reports MAX_MD_DEVS) and\nremove MD_THREAD_NAME_MAX which hasn\u0027t been used for a while.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "850b2b420cd5b363ed4cf48a8816d656c8b5251b", "tree": "b63aa37b3fb7dfb7a0b674d5a2ebe9da13cc7405", "parents": [ "6814d5368d68341ec6b5e4ecd10ea5947130775a" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Tue Oct 03 01:15:46 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Tue Oct 03 08:04:17 2006 -0700" }, "message": "[PATCH] md: replace magic numbers in sb_dirty with well defined bit flags\n\nInstead of magic numbers (0,1,2,3) in sb_dirty, we have\nsome flags instead:\nMD_CHANGE_DEVS\n Some device state has changed requiring superblock update\n on all devices.\nMD_CHANGE_CLEAN\n The array has transitions from \u0027clean\u0027 to \u0027dirty\u0027 or back,\n requiring a superblock update on active devices, but possibly\n not on spares\nMD_CHANGE_PENDING\n A superblock update is underway.\n\nWe wait for an update to complete by waiting for all flags to be clear. A\nflag can be set at any time, even during an update, without risk that the\nchange will be lost.\n\nStop exporting md_update_sb - isn\u0027t needed.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "9361401eb7619c033e2394e4f9f6d410d6719ac7", "tree": "04b94a71f2366988c17740d1c16cfbdec41d5d2e", "parents": [ "d366e40a1cabd453be6e2609caa7e12f9ca17b1f" ], "author": { "name": "David Howells", "email": "dhowells@redhat.com", "time": "Sat Sep 30 20:45:40 2006 +0200" }, "committer": { "name": "Jens Axboe", "email": "axboe@nelson.home.kernel.dk", "time": "Sat Sep 30 20:52:31 2006 +0200" }, "message": "[PATCH] BLOCK: Make it possible to disable the block layer [try #6]\n\nMake it possible to disable the block layer. Not all embedded devices require\nit, some can make do with just JFFS2, NFS, ramfs, etc - none of which require\nthe block layer to be present.\n\nThis patch does the following:\n\n (*) Introduces CONFIG_BLOCK to disable the block layer, buffering and blockdev\n support.\n\n (*) Adds dependencies on CONFIG_BLOCK to any configuration item that controls\n an item that uses the block layer. This includes:\n\n (*) Block I/O tracing.\n\n (*) Disk partition code.\n\n (*) All filesystems that are block based, eg: Ext3, ReiserFS, ISOFS.\n\n (*) The SCSI layer. As far as I can tell, even SCSI chardevs use the\n \t block layer to do scheduling. Some drivers that use SCSI facilities -\n \t such as USB storage - end up disabled indirectly from this.\n\n (*) Various block-based device drivers, such as IDE and the old CDROM\n \t drivers.\n\n (*) MTD blockdev handling and FTL.\n\n (*) JFFS - which uses set_bdev_super(), something it could avoid doing by\n \t taking a leaf out of JFFS2\u0027s book.\n\n (*) Makes most of the contents of linux/blkdev.h, linux/buffer_head.h and\n linux/elevator.h contingent on CONFIG_BLOCK being set. sector_div() is,\n however, still used in places, and so is still available.\n\n (*) Also made contingent are the contents of linux/mpage.h, linux/genhd.h and\n parts of linux/fs.h.\n\n (*) Makes a number of files in fs/ contingent on CONFIG_BLOCK.\n\n (*) Makes mm/bounce.c (bounce buffering) contingent on CONFIG_BLOCK.\n\n (*) set_page_dirty() doesn\u0027t call __set_page_dirty_buffers() if CONFIG_BLOCK\n is not enabled.\n\n (*) fs/no-block.c is created to hold out-of-line stubs and things that are\n required when CONFIG_BLOCK is not set:\n\n (*) Default blockdev file operations (to give error ENODEV on opening).\n\n (*) Makes some /proc changes:\n\n (*) /proc/devices does not list any blockdevs.\n\n (*) /proc/diskstats and /proc/partitions are contingent on CONFIG_BLOCK.\n\n (*) Makes some compat ioctl handling contingent on CONFIG_BLOCK.\n\n (*) If CONFIG_BLOCK is not defined, makes sys_quotactl() return -ENODEV if\n given command other than Q_SYNC or if a special device is specified.\n\n (*) In init/do_mounts.c, no reference is made to the blockdev routines if\n CONFIG_BLOCK is not defined. This does not prohibit NFS roots or JFFS2.\n\n (*) The bdflush, ioprio_set and ioprio_get syscalls can now be absent (return\n error ENOSYS by way of cond_syscall if so).\n\n (*) The seclvl_bd_claim() and seclvl_bd_release() security calls do nothing if\n CONFIG_BLOCK is not set, since they can\u0027t then happen.\n\nSigned-Off-By: David Howells \u003cdhowells@redhat.com\u003e\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n" }, { "commit": "ff4e8d9a9f46e3a7f89d14ade52fe5d53a82c022", "tree": "e15fc4d8af0045a3947e3a25844773c6ee50289a", "parents": [ "0b8c9de05c2a860fe6b02fedcb48763bcee648b3" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jul 10 04:44:16 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Jul 10 13:24:16 2006 -0700" }, "message": "[PATCH] md: fix resync speed calculation for restarted resyncs\n\nWe introduced \u0027io_sectors\u0027 recently so we could count the sectors that causes\nio during resync separate from sectors which didn\u0027t cause IO - there can be a\ndifference if a bitmap is being used to accelerate resync.\n\nHowever when a speed is reported, we find the number of sectors processed\nrecently by subtracting an oldish io_sectors count from a current\n\u0027curr_resync\u0027 count. This is wrong because curr_resync counts all sectors,\nnot just io sectors.\n\nSo, add a field to mddev to store the curren io_sectors separately from\ncurr_resync, and use that in the calculations.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "42543769142d2375f2b5f8fc9cac999f84bd4c4c", "tree": "404606c37da872639111c5b641f07fff2ded1dff", "parents": [ "07d84d109d8beedd68df9da2e4e9f25c8217e7fb" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jun 26 00:27:57 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Jun 26 09:58:39 2006 -0700" }, "message": "[PATCH] md: Don\u0027t write dirty/clean update to spares - leave them alone\n\n- record the \u0027event\u0027 count on each individual device (they\n might sometimes be slightly different now)\n- add a new value for \u0027sb_dirty\u0027: \u00273\u0027 means that the super\n block only needs to be updated to record a clean\u003c-\u003edirty\n transition.\n- Prefer odd event numbers for dirty states and even numbers\n for clean states\n- Using all the above, don\u0027t update the superblock on\n a spare device if the update is just doing a clean-dirty\n transition. To accomodate this, a transition from\n dirty back to clean might now decrement the events counter\n if nothing else has changed.\n\nThe net effect of this is that spare drives will not see any IO requests\nduring normal running of the array, so they can go to sleep if that is what\nthey want to do.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "5fd6c1dce06ec24ef3de20fe0c7ecf2ba9fe5ef9", "tree": "bd7cc8c22615e8622d3d63b7fc68dcc3ac1964de", "parents": [ "a8a55c387da28d67d98f56e4f5021462cb61f7a6" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jun 26 00:27:40 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Jun 26 09:58:37 2006 -0700" }, "message": "[PATCH] md: allow checkpoint of recovery with version-1 superblock\n\nFor a while we have had checkpointing of resync. The version-1 superblock\nallows recovery to be checkpointed as well, and this patch implements that.\n\nDue to early carelessness we need to add a feature flag to signal that the\nrecovery_offset field is in use, otherwise older kernels would assume that a\npartially recovered array is in fact fully recovered.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "8932c2e0dcae52e73430878fd8a7a7800176eada", "tree": "3b734ccaa0fb210f1fd76ff9bb6fb9700d1c4095", "parents": [ "c70810b327a247d12805554d0aa4605e257effb8" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Jun 26 00:27:36 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Jun 26 09:58:36 2006 -0700" }, "message": "[PATCH] md: remove arbitrary limit on chunk size\n\nThe largest chunk size the code can support without substantial surgery is\n2^30 bytes, so make that the limit instead of an arbitrary 4Meg. Some day,\nthe \u0027chunksize\u0027 should change to a sector-shift instead of a byte-count. Then\nno limit would be needed.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "df5b89b323b922f56650b4b4d7c41899b937cf19", "tree": "d1578782a9d763174474fe2db08f4ce4c6dc6218", "parents": [ "48c9c27b8bcd2a328a06151e2d5c1170db0b701b" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 27 01:18:20 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Mar 27 08:45:03 2006 -0800" }, "message": "[PATCH] md: Convert reconfig_sem to reconfig_mutex\n\n... being careful that mutex_trylock is inverted wrt down_trylock\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "e464eafdb4400c6d6576ba3840d8bd40340f8a96", "tree": "0c3f4003c883264ee08300c02007f06e4d1ebb91", "parents": [ "16484bf59634e25d1299761e5ed8bacf22bc6368" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 27 01:18:14 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Mar 27 08:45:02 2006 -0800" }, "message": "[PATCH] md: Support suspending of IO to regions of an md array\n\nThis allows user-space to access data safely. This is needed for raid5\nreshape as user-space needs to take a backup of the first few stripes before\nallowing reshape to commence.\n\nIt will also be useful in cluster-aware raid1 configurations so that all\ncluster members can leave a section of the array untouched while a\nresync/recovery happens.\n\nA \u0027start\u0027 and \u0027end\u0027 of the suspended range are written to 2 sysfs attributes.\nNote that only one range can be suspended at a time.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "63c70c4f3a30e77e6f445bd16eff7934a031ebd3", "tree": "12d2cd6cec5cd304bdf514ec8999a44c26f058ae", "parents": [ "b578d55fdd80140f657130abd85aebeb345755fb" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 27 01:18:13 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Mar 27 08:45:02 2006 -0800" }, "message": "[PATCH] md: Split reshape handler in check_reshape and start_reshape\n\ncheck_reshape checks validity and does things that can be done instantly -\nlike adding devices to raid1. start_reshape initiates a restriping process to\nconvert the whole array.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "f67055780caac6a99f43834795c43acf99eba6a6", "tree": "6b80e7b4cb300edb0910dbad1d840ff8e2f36ae5", "parents": [ "292695531ae4019bb15deedc121b218d1908b648" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 27 01:18:11 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Mar 27 08:45:01 2006 -0800" }, "message": "[PATCH] md: Checkpoint and allow restart of raid5 reshape\n\nWe allow the superblock to record an \u0027old\u0027 and a \u0027new\u0027 geometry, and a\nposition where any conversion is up to. The geometry allows for changing\nchunksize, layout and level as well as number of devices.\n\nWhen using verion-0.90 superblock, we convert the version to 0.91 while the\nconversion is happening so that an old kernel will refuse the assemble the\narray. For version-1, we use a feature bit for the same effect.\n\nWhen starting an array we check for an incomplete reshape and restart the\nreshape process if needed. If the reshape stopped at an awkward time (like\nwhen updating the first stripe) we refuse to assemble the array, and let\nuser-space worry about it.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "ccfcc3c10b2a5cb8fd3c918199a4ff904fc6fb3e", "tree": "0c86fe6ba0f0bafd76f86f81b7425d6d497664aa", "parents": [ "7ecaa1e6a1ad69862e9980b6c777e11f26c4782d" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Mon Mar 27 01:18:09 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Mar 27 08:45:01 2006 -0800" }, "message": "[PATCH] md: Core of raid5 resize process\n\nThis patch provides the core of the resize/expand process.\n\nsync_request notices if a \u0027reshape\u0027 is happening and acts accordingly.\n\nIt allocated new stripe_heads for the next chunk-wide-stripe in the target\ngeometry, marking them STRIPE_EXPANDING.\n\nThen it finds which stripe heads in the old geometry can provide data needed\nby these and marks them STRIPE_EXPAND_SOURCE. This causes stripe_handle to\nread all blocks on those stripes.\n\nOnce all blocks on a STRIPE_EXPAND_SOURCE stripe_head are read, any that are\nneeded are copied into the corresponding STRIPE_EXPANDING stripe_head. Once a\nSTRIPE_EXPANDING stripe_head is full, it is marks STRIPE_EXPAND_READY and then\nis written out and released.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "88202a0c84e1951d6630d1d557d4801a8cc5b5ef", "tree": "5a1aadc008de1ee68f40a4bb667157f248211230", "parents": [ "6d7ff7380b2e28c2807da3bf9fa614d91d15bacf" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 06 00:21:36 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Jan 06 08:34:10 2006 -0800" }, "message": "[PATCH] md: allow sync-speed to be controlled per-device\n\nAlso export current (average) speed and status in sysfs.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nAcked-by: Greg KH \u003cgreg@kroah.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "4dbcdc751cb25ffca3a8374cbc5ab6de961cc545", "tree": "61d1eb6e35c0eec66e5a8610c1f57b1e6eec135c", "parents": [ "d9d166c2a9d5d01af34396793950aa695883eed4" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 06 00:20:52 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Jan 06 08:34:09 2006 -0800" }, "message": "[PATCH] md: count corrected read errors per drive\n\nStore this total in superblock (As appropriate), and make it available to\nuserspace via sysfs.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nAcked-by: Greg KH \u003cgreg@kroah.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "d9d166c2a9d5d01af34396793950aa695883eed4", "tree": "6cb593cc0636476ebec9a9690aa48997510f9701", "parents": [ "8bb93aaca2062cd54cc2c58c76ee8409cae209a7" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 06 00:20:51 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Jan 06 08:34:09 2006 -0800" }, "message": "[PATCH] md: allow array level to be set textually via sysfs\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nAcked-by: Greg KH \u003cgreg@kroah.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "2989ddbd6e1d9638a188311b896362c4bf7b7c25", "tree": "6dd1adb19cdff7135aadb96a309e55aed59e8be9", "parents": [ "f188593ee7af8c71755d2df269a7a5f62c4b695e" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 06 00:20:43 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Jan 06 08:34:07 2006 -0800" }, "message": "[PATCH] md: make a couple of names in md.c static\n\n.. because they aren\u0027t used outside md.c\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "1345b1d8adbdeceb1c871d9a4af5e2a700b341c6", "tree": "0da1db7d273752ad5281db415e9245f3f3a63383", "parents": [ "7dd5d34c6c2da04e8cd0732e08b33cd3359e4bae" ], "author": { "name": "NeilBrown", "email": "neilb@suse.de", "time": "Fri Jan 06 00:20:40 2006 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Fri Jan 06 08:34:07 2006 -0800" }, "message": "[PATCH] md: define and use safe_put_page for md\n\nmd sometimes call put_page on NULL pointers (treating it like kfree). This is\nnot safe, so define and use a \u0027safe_put_page\u0027 which checks for NULL.\n\nSigned-off-by: Neil Brown \u003cneilb@suse.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" } ], "next": "2604b703b6b3db80e3c75ce472a54dfd0b7bf9f4" }