)]}' { "log": [ { "commit": "f6a570333e554b48ad589e7137c77c57809eee81", "tree": "68dd6d5c8fe537a19a84a4189202f5cf70925c17", "parents": [ "2b5f6dcce5bf94b9b119e9ed8d537098ec61c3d2" ], "author": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Wed Oct 18 01:47:25 2006 -0400" }, "committer": { "name": "Al Viro", "email": "viro@zeniv.linux.org.uk", "time": "Mon Dec 04 02:00:22 2006 -0500" }, "message": "[PATCH] severing module.h-\u003esched.h\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\n" }, { "commit": "98644047916c24258fb47c3dab2bed8a44f53b83", "tree": "06b94a000abae5d4710786cc57a5ec424e09cc12", "parents": [ "43a145a3440c5c5f24ff2888801e40e2242187e6" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Mon Oct 16 18:09:39 2006 -0400" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Wed Oct 25 15:13:53 2006 -0700" }, "message": "[SCSI] libiscsi: fix oops in connection create failure path\n\nIf connection creation fails we end up calling list_del\non a invalid struct. This then causes an oops. We are not\nacutally using the lists (old MCS code we thought might\nbe useful elsewhere) so this patch just removes that\ncode.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "60ecebf5a10e42f5e2d6e07eb9e24bdee8500b81", "tree": "79b27fd89fe175daa4e4cd6feda8a6548c6d9bf1", "parents": [ "ffd0436ed2e5a741c8d30062b489b989acf0a526" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Thu Aug 31 18:09:25 2006 -0400" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Sat Sep 02 13:37:07 2006 -0500" }, "message": "[SCSI] add refcouting around ctask usage in main IO patch\n\nIt is possible that a ctask could be completing and getting\ncleaned up at the same time, we are finishing up the last\ndata transfer. This could then result in the data transfer\ncode using stale or invalid values. This patch adds a refcount\nto the ctask. When the count goes to zero then we know the\ntransmit thread and recv thread or softirq are not touching\nit and we can safely release it.\n\nThe eh should not need to grab a reference because it only cleans\nup a task if it has both the xmit mutex and recv lock (or recv\nside suspended).\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "ffd0436ed2e5a741c8d30062b489b989acf0a526", "tree": "037433a76a116c67d3f074c5a83305be8241a8e5", "parents": [ "e5b3cd42960a10c1bc3701d4f00767463c88ec9d" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Thu Aug 31 18:09:24 2006 -0400" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Sat Sep 02 13:37:04 2006 -0500" }, "message": "[SCSI] libiscsi, iscsi_tcp, iscsi_iser: check that burst lengths are valid.\n\niSCSI RFC states that the first burst length must be smaller than the\nmax burst length. We currently assume targets will be good, but that may\nnot be the case, so this patch adds a check.\n\nThis patch also moves the unsol data out offset to the lib so the LLDs\ndo not have to track it.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "c8dc1e523b0f1e6dd71cdabd8c7d7587c6dc27f9", "tree": "466efa148c152f4330c92f02f81bc72e8f395114", "parents": [ "9aaa2b4621280b6de1ecfb6dd7cd5cbe59fd1264" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Mon Jul 24 15:47:39 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Fri Jul 28 11:49:34 2006 -0500" }, "message": "[SCSI] iscsi bugfixes: reduce memory allocations\n\nWe currently try to allocate a max_recv_data_segment_length\nwhich can be very large (default is 64K), and common uses\nare up to 1MB. It is very very difficult to allocte this\nmuch contiguous memory and it turns out we never even use it.\nWe really only need a couple of pages, so this patch has us\nallocates just what we know what we need today.\n\nLater if vendors start adding vendor specific data and\nwe need to handle large buffers we can do this, but for\nthe last 4 years we have not seen anyone do this or request\nit.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "7ea8b82847293c2311cf08fc3ed31ab0e452a27e", "tree": "eab8e9be7cc779cc3c9e48774fcb4b33ebb2c8df", "parents": [ "275fd7d129fdd16d1dc0ec6a8d60bd6b72a76e31" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Mon Jul 24 15:47:22 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Fri Jul 28 11:48:16 2006 -0500" }, "message": "[SCSI] iscsi bugfixes: fix abort handling\n\nAbort handler fixes.\n\nIf a connection is dropped and reconnected while an abort is\nrunning then we should assume the recovery code will clean up\nthe abort. Not doing so causes a oops.\n\nAnd if a command completes then we get the status for the abort, we do not\nneed to call into the LLD to cleanup the resources. Doing this causes\nand oops in iser because it ends up freeing some resources twice.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "b6c395ed0387c824ddf125d3b74b576a2575c149", "tree": "76ca5cd982063335088384622e5033401bbc5057", "parents": [ "d82967c70658a408ea6cae5dc989ba8b2c0999e1" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Mon Jul 24 15:47:15 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Fri Jul 28 11:47:40 2006 -0500" }, "message": "[SCSI] iscsi bugfixes: fix r2t handling\n\nThe iscsi tcp code can pluck multiple rt2s from the tasks\u0027s r2tqueue\nin the xmit code. This can result in the task being queued on the xmit queue\nbut gettting completed at the same time.\n\nThis patch fixes the above bug by making the fifo a list so\nwe always remove the entry on the list del.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "a54a52caad4bd6166cb7fa64e4e93031fa2fda5d", "tree": "728672fae35fd344619129e78213043dabacf099", "parents": [ "01cb225dad8da2e717356fab03240e2f4a8d01bf" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Wed Jun 28 12:00:23 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Thu Jun 29 11:07:14 2006 -0400" }, "message": "[SCSI] iscsi: fixup set/get param functions\n\nReduce duplication in the software iscsi_transport modules by\nadding a libiscsi function to handle the common grunt work.\n\nThis also has the drivers return specifc -EXXX values for different\nerrors so userspace can finally handle them in a sane way.\n\nAlso just pass the sysfs buffers to the drivers so HW iscsi can\nget/set its string values, like targetname, and initiatorname.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "67a611149b2ac5f4af1e36bfffbfe3198cd3712c", "tree": "f74b70aefc16544901a3a3191bee8be8c467747d", "parents": [ "9bf0a28c9a24e2cee5deecf89d118254374c75ba" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Tue May 30 00:37:20 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Mon Jun 05 19:10:45 2006 -0400" }, "message": "[SCSI] iscsi: don\u0027t switch states when just cleaning up\n\nIf recovery failed or we are in recovery only overwrite the state\nif we are going to terminate the session or if we logged back in.\n\nSTOP_CONN_SUSPEND and conn_cnt are not used. We only support\na single connection session ATM, so cleanup that code while\nwe are working around it.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "656cffc95f0cb8211aa75eaca249e6ff4f59ec83", "tree": "7b10e224ef3ac140570101ed98c918949863995b", "parents": [ "790f39a2d5f03623b027f340b945f135d006ceba" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Thu May 18 20:31:42 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Sat May 20 09:36:17 2006 -0500" }, "message": "[SCSI] iscsi: fix command requeues during iscsi recovery\n\nDo not flush queues then block session. This will cause commands\nto needlessly swing around on us and remove goofy\nrecovery_failed field and replace with state value.\n\nAnd do not start recovery from within the host reset function.\nThis causeis too many problems becuase open-iscsi was desinged to\ncall out to userspace then have userpscae decide if we should\ngo into recovery or kill the session.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "be2df72e7ec5fa5e6e1ccccab6cef97ecbb9c191", "tree": "f4a78a36be0fac5304960e615ad46cc127cf6306", "parents": [ "ed2abc7ff19dc99c6242a70f8578a17b2ff0d0ce" ], "author": { "name": "Or Gerlitz", "email": "ogerlitz@voltaire.com", "time": "Tue May 02 19:46:43 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Wed May 10 10:12:52 2006 -0500" }, "message": "[SCSI] iscsi: align printks\n\nalign printk output\n\nSigned-off-by: Or Gerlitz \u003cogerlitz@voltaire.com\u003e\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" }, { "commit": "7996a778ff8c717cb1a7a294475c59cc8f1e9fb8", "tree": "3d7ee67ad547a65ad10f5c7e41e20f6124ee249b", "parents": [ "30a6c65236f9d26e3325cae468f330b833a3878c" ], "author": { "name": "Mike Christie", "email": "michaelc@cs.wisc.edu", "time": "Thu Apr 06 21:13:41 2006 -0500" }, "committer": { "name": "James Bottomley", "email": "jejb@mulgrave.il.steeleye.com", "time": "Fri Apr 14 14:09:00 2006 -0500" }, "message": "[SCSI] iscsi: add libiscsi\n\nThere is a lot of code duplcited between iscsi_tcp\nand the upcoming iscsi_iser driver. This patch puts\nthe duplicated code in a lib. There is more code\nto move around but this takes care of the\nbasics. For iscsi_offload if they use the lib we will\nprobably move some things around. For example in the\nqueuecommand we will not assume that the LLD wants\nto do queue_work, but it is better to handle that\nlater when we know for sure what iscsi_offload looks\nlike (we could probably do this for iscsi_iser though to).\n\nIdeally I would like to get the iscsi_transports modules\nto a place where all they really have to do is put data\non the wire, but how to do that will hopefully be more clear\nwhen we see other modules like iscsi_offload. Or maybe\niscsi_offload will not use the lib and it will just be\niscsi_iser and iscsi_tcp and maybe the iscsi_tcp_tgt if that\nis allowed in mainline.\n\nSigned-off-by: Mike Christie \u003cmichaelc@cs.wisc.edu\u003e\nSigned-off-by: James Bottomley \u003cJames.Bottomley@SteelEye.com\u003e\n" } ] }