)]}' { "log": [ { "commit": "4f5058c3b71ed5930bb2b478c4d5dbc799dd9ad1", "tree": "5c756cbcd329d62fb95744185b54c952eed903df", "parents": [ "78fca1b95837a44983608302e89c7258981be9b3" ], "author": { "name": "Xiaotian Feng", "email": "dfeng@redhat.com", "time": "Sat Apr 02 19:39:35 2011 +0800" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Apr 02 21:26:20 2011 +0200" }, "message": "genirq: Fix cpumask leak in __setup_irq()\n\nThe allocated cpumask should be freed in __setup_irq().\n\nSigned-off-by: Xiaotian Feng \u003cdfeng@redhat.com\u003e\nLKML-Reference: \u003c1301744375-6812-1-git-send-email-dfeng@redhat.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "0c6f8a8b917ad361319c8ace3e9f28e69bfdb4c1", "tree": "b4b0cb4b619368bc93ff883f4b667e05a185549b", "parents": [ "dced35aeb0367dda2636ee9ee914bda14510dcc9" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 28 13:32:20 2011 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 29 14:48:19 2011 +0200" }, "message": "genirq: Remove compat code\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "cd22c0e44b105aecd78e5f9e77abab3a1b8dc00c", "tree": "98be0a3bdb0fcc9ff7f31e774a4c69d655d19041", "parents": [ "89078d572eb9ce8d4c04264b8b0ba86de0d74c8f" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 29 11:36:05 2011 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 29 11:36:05 2011 +0200" }, "message": "genirq: Fix harmless typo\n\nThe late night fixup missed to convert the data type from irq_desc to\nirq_data, which results in a harmless but annoying warning.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "0ef5ca1e1f0de71300142b8f730f26ded6a0c2f3", "tree": "f8eb0e5e63b7d3115c27f110e59a762fd871d693", "parents": [ "a6aeddd1c4e464a2150f97ca2d1c3d68cfbd9296" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 28 21:59:37 2011 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 29 01:41:22 2011 +0200" }, "message": "genirq; Fix cleanup fallout\n\nI missed the CONFIG_GENERIC_PENDING_IRQ dependency in the affinity\nrelated functions and the IRQ_LEVEL propagation into irq_data\nstate. Did not pop up on my main test platforms. :(\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nTested-by: David Daney \u003cddaney@caviumnetworks.com\u003e\n" }, { "commit": "30398bf6c684a77274dbdabf7efc1f24e4a99028", "tree": "d4ee5d514a53fbb8186bdbc726a525bb488c7dcf", "parents": [ "ee38c04b58983a236b43bae71b2415d38bceaf75" ], "author": { "name": "Randy Dunlap", "email": "randy.dunlap@oracle.com", "time": "Fri Mar 18 09:33:56 2011 -0700" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 28 20:13:57 2011 +0200" }, "message": "genirq: Fix new kernel-doc warnings\n\nFix new irq-related kernel-doc warnings in 2.6.38:\n\nWarning(kernel/irq/manage.c:149): No description found for parameter \u0027mask\u0027\nWarning(kernel/irq/manage.c:149): Excess function parameter \u0027cpumask\u0027 description in \u0027irq_set_affinity\u0027\nWarning(include/linux/irq.h:161): No description found for parameter \u0027state_use_accessors\u0027\nWarning(include/linux/irq.h:161): Excess struct/union/enum/typedef member \u0027state_use_accessor\u0027 description in \u0027irq_data\u0027\n\nSigned-off-by: Randy Dunlap \u003crandy.dunlap@oracle.com\u003e\nLKML-Reference: \u003c20110318093356.b939558d.randy.dunlap@oracle.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "32f4125ebffee4f3c4dbc6a437fc656129eb9e60", "tree": "d64c6bb7ba40c33734896303734416ea5b4f3290", "parents": [ "c2d0c555c22242c3a76e366074c4d83ef9fa3b8c" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 28 14:10:52 2011 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 28 16:55:10 2011 +0200" }, "message": "genirq: Move INPROGRESS, MASKED and DISABLED state flags to irq_data\n\nWe really need these flags for some of the interrupt chips. Move it\nfrom internal state to irq_data and provide proper accessors.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: David Daney \u003cddaney@caviumnetworks.com\u003e\n" }, { "commit": "c2d0c555c22242c3a76e366074c4d83ef9fa3b8c", "tree": "7426945b566129163639229fa0a5f420cdce13a8", "parents": [ "b3d422329f2e061d66af4f933ef316e50e5edcac" ], "author": { "name": "David Daney", "email": "ddaney@caviumnetworks.com", "time": "Fri Mar 25 12:38:50 2011 -0700" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sun Mar 27 17:45:59 2011 +0200" }, "message": "genirq: Split irq_set_affinity() so it can be called with lock held.\n\nThe .irq_cpu_online() and .irq_cpu_offline() functions may need to\nadjust affinity, but they are called with the descriptor lock held.\nCreate __irq_set_affinity_locked() which is called with the lock held.\nMake irq_set_affinity() just a wrapper that acquires the lock.\n\n[ tglx: Changed the argument to irq_data, added a !desc check and\n moved the !irq_set_affinity check where it belongs ]\n\nSigned-off-by: David Daney \u003cddaney@caviumnetworks.com\u003e\nCc: linux-mips@linux-mips.org\nCc: ralf@linux-mips.org\nLKML-Reference: \u003c1301081931-11240-4-git-send-email-ddaney@caviumnetworks.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "1c389795c15d349c2c7b23baf927e16e86ca3ae8", "tree": "cc06eb6e38ad3cd291e9420ec5203a89c77bf3cc", "parents": [ "15825a5cd4adb199fd76737e8bc846154e315218" ], "author": { "name": "Dan Carpenter", "email": "error27@gmail.com", "time": "Thu Mar 17 14:43:07 2011 +0300" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Mar 17 15:52:30 2011 +0100" }, "message": "genirq: Fix incorrect unlock in __setup_irq()\n\ngoto out_thread is called before we take the lock. It causes a gcc\nwarning: \"kernel/irq/manage.c:858: warning: ‘flags’ may be used\nuninitialized in this function\"\n\n[ tglx: Moved unlock before free_cpumask_var() ]\n\nSigned-off-by: Dan Carpenter \u003cerror27@gmail.com\u003e\nLKML-Reference: \u003c20110317114307.GJ2008@bicker\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "8d32a307e4faa8b123dc8a9cd56d1a7525f69ad3", "tree": "50141928db9f0ac1377dd8bb355d600232d6b38c", "parents": [ "8eb90c30e0e815a1308828352eabd03ca04229dd" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Feb 23 23:52:23 2011 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 26 11:57:18 2011 +0100" }, "message": "genirq: Provide forced interrupt threading\n\nAdd a commandline parameter \"threadirqs\" which forces all interrupts except\nthose marked IRQF_NO_THREAD to run threaded. That\u0027s mostly a debug option to\nallow retrieving better debug data from crashing interrupt handlers. If\n\"threadirqs\" is not enabled on the kernel command line, then there is no\nimpact in the interrupt hotpath.\n\nArchitecture code needs to select CONFIG_IRQ_FORCED_THREADING after\nmarking the interrupts which cant be threaded IRQF_NO_THREAD. All\ninterrupts which have IRQF_TIMER set are implict marked\nIRQF_NO_THREAD. Also all PER_CPU interrupts are excluded.\n\nForced threading hard interrupts also forces all soft interrupt\nhandling into thread context.\n\nWhen enabled it might slow down things a bit, but for debugging problems in\ninterrupt code it\u0027s a reasonable penalty as it does not immediately\ncrash and burn the machine when an interrupt handler is buggy.\n\nSome test results on a Core2Duo machine:\n\nCache cold run of:\n # time git grep irq_desc\n\n non-threaded threaded\n real 1m18.741s 1m19.061s\n user 0m1.874s 0m1.757s\n sys 0m5.843s 0m5.427s\n\n # iperf -c server\nnon-threaded\n[ 3] 0.0-10.0 sec 1.09 GBytes 933 Mbits/sec\n[ 3] 0.0-10.0 sec 1.09 GBytes 934 Mbits/sec\n[ 3] 0.0-10.0 sec 1.09 GBytes 933 Mbits/sec\nthreaded\n[ 3] 0.0-10.0 sec 1.09 GBytes 939 Mbits/sec\n[ 3] 0.0-10.0 sec 1.09 GBytes 934 Mbits/sec\n[ 3] 0.0-10.0 sec 1.09 GBytes 937 Mbits/sec\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20110223234956.772668648@linutronix.de\u003e\n" }, { "commit": "9d591edd02a245305b1b9379e4c5571bad4d2774", "tree": "1fdd608db6096afaf2696ff9878cb4b7096047f3", "parents": [ "b5faba21a6805c33b40e258d36f57997ee1de131" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Feb 23 23:52:16 2011 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 25 20:24:21 2011 +0100" }, "message": "genirq: Allow shared oneshot interrupts\n\nSupport ONESHOT on shared interrupts, if all drivers agree on it.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20110223234956.483640430@linutronix.de\u003e\n" }, { "commit": "b5faba21a6805c33b40e258d36f57997ee1de131", "tree": "c84ef3357ecd6e1b1cfda623136529db0e5fab6f", "parents": [ "1204e95689f9fbd245a4ce5c1b0cd0a9b77f8d25" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Feb 23 23:52:13 2011 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 25 20:24:21 2011 +0100" }, "message": "genirq: Prepare the handling of shared oneshot interrupts\n\nFor level type interrupts we need to track how many threads are on\nflight to avoid useless interrupt storms when not all thread handlers\nhave finished yet. Keep track of the woken threads and only unmask\nwhen there are no more threads in flight.\n\nYes, I\u0027m lazy and using a bitfield. But not only because I\u0027m lazy, the\nmain reason is that it\u0027s way simpler than using a refcount. A refcount\nbased solution would need to keep track of various things like\ncrashing the irq thread, spurious interrupts coming in,\ndisables/enables, free_irq() and some more. The bitfield keeps the\ntracking simple and makes things just work. It\u0027s also nicely confined\nto the thread code pathes and does not require additional checks all\nover the place.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20110223234956.388095876@linutronix.de\u003e\n" }, { "commit": "8fff39e06987492da3d4a0b9ec7cdbd245b6762b", "tree": "751d1930d88def0238e927ac4e03f23703656a4d", "parents": [ "ed4dea6e0e33a3e58d8b77b775a8f0e433e7a005" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 21 14:19:42 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 21 21:20:00 2011 +0100" }, "message": "genirq: Add missing break in __irq_set_trigger()\n\nThe switch case in __irq_set_trigger() lacks a break, which emits a\npr_err unconditionally on success.\n\nReported-by: Lars-Peter Clausen \u003clars@metafoo.de\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "02725e7471b8dd58fa96f6604bdb5dde45405a2e", "tree": "7c5243d0a2e33f2165d1531febbe9752fad389c7", "parents": [ "d5eb4ad2dfb2dfae43fd51bc8630b4fc3ef00e92" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 12 10:37:36 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:23 2011 +0100" }, "message": "genirq: Use irq_get/put functions\n\nConvert the management functions to use the common irq_get/put\nfunction.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "091738a266fc74329ae186f22ff2b3f01319112d", "tree": "75190438b74ad345753a2a8c239724995cee40ea", "parents": [ "a6967caf00ebbb2d4acdebcb72a25f2e9ba43fd2" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 14 20:16:43 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:23 2011 +0100" }, "message": "genirq: Remove real old transition functions\n\nThese transition helpers are stale for years now. Remove them.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "7f94226f03299f1ca32f118f02f2a0295e0e5e93", "tree": "049a83a3c264d4378f341991800c345bf4eeb5a5", "parents": [ "d4d5e08960844a062da8387ee5f16ca7a33200d0" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 10 19:46:26 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:22 2011 +0100" }, "message": "genirq: Move wakeup state to irq_data\n\nSome irq_chips need to know the state of wakeup mode for\nsetting the trigger type etc. Reflect it in irq_data state.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "d4d5e08960844a062da8387ee5f16ca7a33200d0", "tree": "1154fd87e25595ae9391740a1d9a3d60f69770de", "parents": [ "2bff17ad2107c66fc8ca96501a7128dd7fa7a390" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 10 13:16:14 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:22 2011 +0100" }, "message": "genirq: Add IRQCHIP_SET_TYPE_MASKED flag\n\nirq_chips, which require to mask the chip before changing the trigger\ntype should set this flag. So the core takes care of it and the\nrequirement for looking into desc-\u003estatus in the chip goes away.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Linus Walleij \u003clinus.walleij@stericsson.com\u003e\nCc: Lars-Peter Clausen \u003clars@metafoo.de\u003e\n" }, { "commit": "1ccb4e612f68ceefb888c2c6c1def6294ea8666d", "tree": "f3cae56489a56ebc54e06871abdf63193dc5a4bb", "parents": [ "876dbd4cc1b35c1a4cb96a2be1d43ea0eabce3b4" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Feb 09 14:44:17 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:21 2011 +0100" }, "message": "genirq: Wrap the remaning IRQ_* flags\n\nUse wrappers to keep them away from the core code.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "876dbd4cc1b35c1a4cb96a2be1d43ea0eabce3b4", "tree": "9be1e7e4cd4a4c9fadd98a9ac637020417215521", "parents": [ "2bdd10558c8d93009cb6c32ce9e30800fbb08add" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 17:28:12 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:20 2011 +0100" }, "message": "genirq: Mirror irq trigger type bits in irq_data.state\n\nThat\u0027s the data structure chip functions get provided. Also allow them\nto signal the core code that they updated the flags in irq_data.state\nby returning IRQ_SET_MASK_OK_NOCOPY. The default is unchanged.\n\nThe type bits should be accessed via:\n\nval \u003d irqd_get_trigger_type(irqdata);\nand\nirqd_set_trigger_type(irqdata, val);\n\nCoders who access them directly will be tracked down and slapped with\nstinking trouts.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "2bdd10558c8d93009cb6c32ce9e30800fbb08add", "tree": "f25b36c02faa8709628e2dbb93748f370678bb52", "parents": [ "bce43032ad79fae0ce5b6174ce1321e643ceb54b" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 17:22:00 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:20 2011 +0100" }, "message": "genirq: Move IRQ_AFFINITY_SET to core\n\nKeep status in sync until last abuser is gone.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "bce43032ad79fae0ce5b6174ce1321e643ceb54b", "tree": "1a4725216ef2d008cbb6a23966785b58088ebf93", "parents": [ "a005677b3dd05decdd8880cf3044ae709856f58f" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 10 22:37:41 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:20 2011 +0100" }, "message": "genirq: Reuse existing can set affinty check\n\nAdd a !desc check while at it.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "a005677b3dd05decdd8880cf3044ae709856f58f", "tree": "529d1454940fe8c6723bd54f01e009d9be2ab840", "parents": [ "1ce6068dac1924f7095be5850481e790cbf1b3c1" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 17:11:03 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:20 2011 +0100" }, "message": "genirq: Mirror IRQ_PER_CPU and IRQ_NO_BALANCING in irq_data.state\n\nThat\u0027s the right data structure to look at for arch code.\n\nAccessor functions are provided.\n\n\t irqd_is_per_cpu(irqdata);\n\t irqd_can_balance(irqdata);\n\nCoders who access them directly will be tracked down and slapped with\nstinking trouts.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "fae581e588e64a0690f3fc995e404fcacaebe772", "tree": "921a408384bf09834662a8e13af08eee87b66d30", "parents": [ "8f53f92404bead2ab2154d45c8f508880bb5d95d" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 16:53:24 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:19 2011 +0100" }, "message": "genirq: Remove CHECK_IRQ_PER_CPU from core code\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "6a58fb3bad099076f36f0f30f44507bc3275cdb6", "tree": "9a40117c941d914b1ef13463436657ae9721fea3", "parents": [ "f230b6d5c48f8d12f4dfa1f8b5ab0b0320076d21" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 15:40:05 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:19 2011 +0100" }, "message": "genirq: Remove CONFIG_IRQ_PER_CPU\n\nThe saving of this switch is minimal versus the ifdef mess it\ncreates. Simple enable PER_CPU unconditionally and remove the config\nswitch.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "f230b6d5c48f8d12f4dfa1f8b5ab0b0320076d21", "tree": "d4d30f77ab196930dd2d2982e90b1f2b5fd9bbe2", "parents": [ "91c499178139d6597e68db19638e4135510a34b8" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 05 15:20:04 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:18 2011 +0100" }, "message": "genirq: Add IRQ_MOVE_PENDING to irq_data.state\n\nchip implementations need to know about it. Keep status in sync until\nall users are fixed. \n\nAccessor function: irqd_is_setaffinity_pending(irqdata)\n\nCoders who access them directly will be tracked down and slapped with\nstinking trouts.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "6d2cd17fde1fc3e93302815f049f255bb2b3123e", "tree": "48995b27264ec657cae21e22a5be1f4c05973bc0", "parents": [ "c531e8361f1968d664e6e97fbd3bfa4cf0e62e42" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 14:34:18 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:18 2011 +0100" }, "message": "genirq: Move IRQ_WAKEUP to core\n\nNo users outside of core.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "c531e8361f1968d664e6e97fbd3bfa4cf0e62e42", "tree": "45b5dbd85203033af684a9588224044289236b0c", "parents": [ "6e40262ea43c4b0e3f435b3a083e4461ef921c17" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 12:44:58 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:18 2011 +0100" }, "message": "genirq: Move IRQ_SUSPENDED to core\n\nNo users outside of core.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "6e40262ea43c4b0e3f435b3a083e4461ef921c17", "tree": "4548c1607f81701b77c0d63b544096cdee8ccde0", "parents": [ "2a0d6fb335d4428285dab2d254911748e6040807" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 12:36:06 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:17 2011 +0100" }, "message": "genirq: Move IRQ_MASKED to core\n\nKeep status in sync until all users are fixed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "2a0d6fb335d4428285dab2d254911748e6040807", "tree": "16090086899df6886436b89cc76e6d0d5b36d083", "parents": [ "c1594b77e46124bb462f961e536120e471c67446" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 12:17:57 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:17 2011 +0100" }, "message": "genirq: Move IRQ_PENDING flag to core\n\nKeep status in sync until all users are fixed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "c1594b77e46124bb462f961e536120e471c67446", "tree": "3f52ad5809125ab5be6db1fd4b1212fe6127df66", "parents": [ "163ef3091195f514a06f064b12914597d2644c55" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 22:11:30 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:17 2011 +0100" }, "message": "genirq: Move IRQ_DISABLED to core\n\nKeep status in sync until all abusers are fixed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "163ef3091195f514a06f064b12914597d2644c55", "tree": "a6e222b1a7366aa93c884257063aa36ef17cf91f", "parents": [ "3d67baec7f1b01fc289ac1a2f1a7e6d5e43391c6" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 11:39:15 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:16 2011 +0100" }, "message": "genirq: Move IRQ_REPLAY and IRQ_WAITING to core\n\nNo users outside of core.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "3d67baec7f1b01fc289ac1a2f1a7e6d5e43391c6", "tree": "f496998278faa653174e185924acb0c74ac2e1b2", "parents": [ "009b4c3b8ad584b3462734127a5bec680d5d6af4" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 21:02:10 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:16 2011 +0100" }, "message": "genirq: Move IRQ_ONESHOT to core\n\nNo users outside of core.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "009b4c3b8ad584b3462734127a5bec680d5d6af4", "tree": "f7cb860dde86ba96afcc085ec7a75ca2a41e49b6", "parents": [ "6954b75b488dd740950573f244ddd66fd28620aa" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 21:48:49 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:16 2011 +0100" }, "message": "genirq: Add IRQ_INPROGRESS to core\n\nWe need to maintain the flag for now in both fields status and istate.\nAdd a CONFIG_GENERIC_HARDIRQS_NO_COMPAT switch to allow testing w/o\nthe status one. Wrap the access to status IRQ_INPROGRESS in a inline\nwhich can be turned of with CONFIG_GENERIC_HARDIRQS_NO_COMPAT along\nwith the define.\n\nThere is no reason that anything outside of core looks at this. That\nneeds some modifications, but we\u0027ll get there.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "7acdd53e5b2c55b6f7e3427e85e2f91fa814a4f9", "tree": "4806536bae32605a3107d6c4384547693eda566c", "parents": [ "bd062e7667ac173afef57fbfe9327f3b914a9d4c" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 20:40:54 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:15 2011 +0100" }, "message": "genirq: Move IRQ_SPURIOUS_DISABLED to core state\n\nNo users outside.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "bd062e7667ac173afef57fbfe9327f3b914a9d4c", "tree": "29e798cb397311abec53ad45b73efdede1fc37b7", "parents": [ "e6bea9c404699223322d7411c6f2ceaec02fa83c" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 20:25:25 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:15 2011 +0100" }, "message": "genirq: Move IRQ_AUTODETECT to internal state\n\nNo users outside of core\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "35e857cbeb24e75c6f9a9312ac30454eee8c5950", "tree": "3f420843de7a59adf6a9f501980e00b45d8a7e61", "parents": [ "c78b9b65faa291def628dbd8539649f58299f0f3" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 10 12:20:23 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:13 2011 +0100" }, "message": "genirq: Fixup core code namespace fallout\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "3aae994fb0f43f6d94a31c33536a83869504abdf", "tree": "f928d2067cecfbb89c1133a8b513f84c890f7e9d", "parents": [ "50f7c0327513d5acefbe26fd33498af18d1ffac5" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 04 10:17:52 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:11 2011 +0100" }, "message": "genirq: Consolidate IRQ_DISABLED\n\nHandle IRQ_DISABLED consistent.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "50f7c0327513d5acefbe26fd33498af18d1ffac5", "tree": "cab9352b720cccbe0b230b97b26b2bc40ca0405b", "parents": [ "87923470c712dff00b101ffb6b6fbc27bd7a6df5" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 03 13:23:54 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:10 2011 +0100" }, "message": "genirq: Remove default magic\n\nNow that everything uses the wrappers, we can remove the default\nfunctions. None of those functions is performance critical.\n\nThat makes the IRQ_MASKED flag tracking fully consistent.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "87923470c712dff00b101ffb6b6fbc27bd7a6df5", "tree": "35dc1a0adec4d00bbc872ce096c481cf0d69314a", "parents": [ "4699923861513671d3f6ade8efb4e56a9a7ecadf" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 03 12:27:44 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:10 2011 +0100" }, "message": "genirq: Consolidate disable/enable\n\nCreate irq_disable/enable and use them to keep the flags consistent.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "4699923861513671d3f6ade8efb4e56a9a7ecadf", "tree": "862cda99b9af6fe96fda955107f24a7b5cc97a18", "parents": [ "3b56f0585fd4c02d047dc406668cb40159b2d340" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Feb 02 21:41:14 2011 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:10 2011 +0100" }, "message": "genirq: Consolidate startup/shutdown of interrupts\n\nAside of duplicated code some of the startup/shutdown sites do not\nhandle the MASKED/DISABLED flags and the depth field at all. Move that\nto a helper function and take care of it there.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20110202212551.787481468@linutronix.de\u003e\n" }, { "commit": "3b56f0585fd4c02d047dc406668cb40159b2d340", "tree": "99ac7c86428aeeaf4d8af51aad5753ac2e6d3af6", "parents": [ "1535dfacbf21c4da1b73fcf07c39913da5bd5581" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Feb 02 21:41:12 2011 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:10 2011 +0100" }, "message": "genirq: Remove bogus conditional\n\nThe if (chip-\u003eirq_shutdown) check will always evaluate to true, as we\nfill in chip-\u003eirq_shutdown with default_shutdown in\nirq_chip_set_defaults() if the chip does not provide its own function.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20110202212551.667607458@linutronix.de\u003e\n" }, { "commit": "2b879eaf095878430c38cbd95e5c0fc4ce65ad8e", "tree": "2459e581520f2dd64f7a9ab75306b2af049d1362", "parents": [ "3b8249e759c701c4a82f99d957be651a7657bf6f" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 14 11:25:02 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:08 2011 +0100" }, "message": "genirq: Remove redundant thread affinity setting\n\nThread affinity is already set by setup_affinity().\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "3b8249e759c701c4a82f99d957be651a7657bf6f", "tree": "f758675da3bb37282eefd50e57456d09b208b452", "parents": [ "569bda8df11effa03e618729293c7961696abb10" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 16:02:20 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:07 2011 +0100" }, "message": "genirq: Do not copy affinity before set\n\nWhile rumaging through arch code I found that there are a few\nworkarounds which deal with the fact that the initial affinity setting\nfrom request_irq() copies the mask into irq_data-\u003eaffinity before the\nchip code is called. In the normal path we unconditionally copy the\nmask when the chip code returns 0.\n\nCopy after the code is called and add a return code\nIRQ_SET_MASK_OK_NOCOPY for the chip functions, which prevents the\ncopy. That way we see the real mask when the chip function decided to\ntruncate it further as some arches do. IRQ_SET_MASK_OK is 0, which is\nthe current behaviour.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "569bda8df11effa03e618729293c7961696abb10", "tree": "b0aed74a4b00a481738dc76a739f6f67e09a99c1", "parents": [ "b008207cbd0d5ce606a1a2ac52826e0ab37d0b99" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 17:05:08 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:07 2011 +0100" }, "message": "genirq: Always apply cpu online mask\n\nIf the affinity had been set by the user, then a later request_irq()\nwill honour that setting. But online cpus can have changed. So apply\nthe online mask and for this case as well.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "b008207cbd0d5ce606a1a2ac52826e0ab37d0b99", "tree": "ef3d7bc8414aa232829483e64f37758f3fdeebc2", "parents": [ "1fa46f1f070961783661ae640cd2f6b2557f3885" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 17:30:50 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:07 2011 +0100" }, "message": "genirq: Rremove redundant check\n\nIRQ_NO_BALANCING is already checked in irq_can_set_affinity() above,\nno need to check it again.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "1fa46f1f070961783661ae640cd2f6b2557f3885", "tree": "cef8928b1a89edc251f8b3f9d9a67bfaa10248af", "parents": [ "a0cd9ca2b907d7ee26575e7b63ac92dad768a75e" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Feb 07 16:46:58 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:07 2011 +0100" }, "message": "genirq: Simplify affinity related code\n\nThere is lot of #ifdef CONFIG_GENERIC_PENDING_IRQ along with\nduplicated code in the irq core. Move the #ifdeffery into one place\nand cleanup the code so it\u0027s readable. No functional change.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "a0cd9ca2b907d7ee26575e7b63ac92dad768a75e", "tree": "4b46f7951b9cd76531caf6d4d68d687e1b622336", "parents": [ "43abe43ce0619d744c7a5bb15cce075e532b53b7" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Feb 10 11:36:33 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:06 2011 +0100" }, "message": "genirq: Namespace cleanup\n\nThe irq namespace has become quite convoluted. My bad. Clean it up\nand deprecate the old functions. All new functions follow the scheme:\n\nirq number based:\n irq_set/get/xxx/_xxx(unsigned int irq, ...)\n\nirq_data based:\n\t irq_data_set/get/xxx/_xxx(struct irq_data *d, ....)\n\nirq_desc based:\n\t irq_desc_get_xxx(struct irq_desc *desc)\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "43abe43ce0619d744c7a5bb15cce075e532b53b7", "tree": "f915103a0d1d671d2f537c7852869b7ca4a0078b", "parents": [ "e7bcecb7b1d29b9ad5af939149a945658620ca8f" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 12 12:10:49 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:58:06 2011 +0100" }, "message": "genirq: Add missing buslock to set_irq_type(), set_irq_wake()\n\nchips behind a slow bus cannot update the chip under desc-\u003elock, but\nwe miss the chip_buslock/chip_bus_sync_unlock() calls around the set\ntype and set wake functions.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "218502bfe674f570205367b9094048207b04ba15", "tree": "c6187b97e7c79d902aef08f049e9fba04c421d56", "parents": [ "51327ada7142ab520ed610a42572d1f4cbfbb2dc", "6d83f94db95cfe65d2a6359cccdf61cf087c2598" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:56:36 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:56:43 2011 +0100" }, "message": "Merge branch \u0027irq/urgent\u0027 into irq/core\n\nReason: Further patches are conflicting with mainline fixes\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "6d83f94db95cfe65d2a6359cccdf61cf087c2598", "tree": "e685c8d43da6dec24901658905cf3cf08e127633", "parents": [ "c1ee6264280e740a9d3ff3feef38642cf0a57013" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 18 23:27:23 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Feb 19 12:11:13 2011 +0100" }, "message": "genirq: Disable the SHIRQ_DEBUG call in request_threaded_irq for now\n\nWith CONFIG_SHIRQ_DEBUG\u003dy we call a newly installed interrupt handler\nin request_threaded_irq().\n\nThe original implementation (commit a304e1b8) called the handler\n_BEFORE_ it was installed, but that caused problems with handlers\ncalling disable_irq_nosync(). See commit 377bf1e4.\n\nIt\u0027s braindead in the first place to call disable_irq_nosync in shared\nhandlers, but ....\n\nMoving this call after we installed the handler looks innocent, but it\nis very subtle broken on SMP.\n\nInterrupt handlers rely on the fact, that the irq core prevents\nreentrancy.\n\nNow this debug call violates that promise because we run the handler\nw/o the IRQ_INPROGRESS protection - which we cannot apply here because\nthat would result in a possibly forever masked interrupt line.\n\nA concurrent real hardware interrupt on a different CPU results in\nhandler reentrancy and can lead to complete wreckage, which was\nunfortunately observed in reality and took a fricking long time to\ndebug.\n\nLeave the code here for now. We want this debug feature, but that\u0027s\nnot easy to fix. We really should get rid of those\ndisable_irq_nosync() abusers and remove that function completely.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Anton Vorontsov \u003cavorontsov@ru.mvista.com\u003e\nCc: David Woodhouse \u003cdwmw2@infradead.org\u003e\nCc: Arjan van de Ven \u003carjan@infradead.org\u003e\nCc: stable@kernel.org # .28 -\u003e .37\n" }, { "commit": "51327ada7142ab520ed610a42572d1f4cbfbb2dc", "tree": "3dd587e6322073d0f8847ef92a835fd9fe456b49", "parents": [ "44951a60ff888add9e84f509ffce20052e45af94", "986c011ddbb3ed44b35e1bfd67f6aa60b293b495" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 11 00:26:54 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 11 00:27:03 2011 +0100" }, "message": "Merge branch \u0027irq/for-mips\u0027 into irq/core\n\nReason: irq/for-mips is provided for mips to make core independent\n progress. Merge it into irq/core to avoid conflicts\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "986c011ddbb3ed44b35e1bfd67f6aa60b293b495", "tree": "defe8664b0287dcda9a7c0767b629eeae0504519", "parents": [ "100b33c8bd8a3235fd0b7948338d6cbb3db3c63d" ], "author": { "name": "David Daney", "email": "ddaney@caviumnetworks.com", "time": "Wed Feb 09 16:04:25 2011 -0800" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 11 00:25:13 2011 +0100" }, "message": "genirq: Call bus_lock/unlock functions in setup_irq()\n\nirq_chips that supply .irq_bus_lock/.irq_bus_sync_unlock functions,\nexpect that the other chip methods will be called inside of calls to\nthe pair. If this expectation is not met, things tend to not work.\n\nMake setup_irq() call chip_bus_lock()/chip_bus_sync_unlock() too.\n\nFor the vast majority of irq_chips, this will be a NOP as most don\u0027t\nhave these bus lock functions.\n\n[ tglx: No we don\u0027t want to call that in __setup_irq(). Way too many\n \terror exit pathes. ]\n\nSigned-off-by: David Daney \u003cddaney@caviumnetworks.com\u003e\nLKML-Reference: \u003c1297296265-18680-1-git-send-email-ddaney@caviumnetworks.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "c9a443cdf7726ce8b78c3177c6ae601ce37292fc", "tree": "4191db70a690b8eb7c86d8c3c10f827e4f2ab32a", "parents": [ "285c1a2c3a5f84ce1c811ab4cb1f8a17466e1a06", "dc5f219e88294b93009eef946251251ffffb6d60" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 16:38:00 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 16:38:21 2011 +0100" }, "message": "Merge branch \u0027irq/for-xen\u0027 into irq/core\n\nirq/for-xen contains new functionality to avoid Xen private irq\nhackery. That branch has a single irq commit and is pulled by Xen to\nbase their new features on.\n\nMerge it into irq/core as other patches modify the same code.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "dc5f219e88294b93009eef946251251ffffb6d60", "tree": "d4fd68261803ee6bd90ed4b422053d879352228e", "parents": [ "100b33c8bd8a3235fd0b7948338d6cbb3db3c63d" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Feb 04 13:19:20 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Feb 08 16:36:47 2011 +0100" }, "message": "genirq: Add IRQF_FORCE_RESUME\n\nXen needs to reenable interrupts which are marked IRQF_NO_SUSPEND in the\nresume path. Add a flag to force the reenabling in the resume code.\n\nTested-and-acked-by: Ian Campbell \u003cIan.Campbell@eu.citrix.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "1fb0ef31f428f345a7c3666f8e7444a563edd537", "tree": "aef752257568dfc69e76902750937514cac95cf0", "parents": [ "cd7eab44e9946c28d595abe3e9a43e945bc49141" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Jan 31 08:57:41 2011 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Jan 31 08:57:41 2011 +0100" }, "message": "genirq: Fix affinity notifier fallout\n\nThe new code of commit cd7eab44e(genirq: Add IRQ affinity notifiers)\nreferences irq_desc.affinity which fails to compile with\nCONFIG_GENERIC_HARDIRQS_NO_DEPRECATED\u003dy.\n\nUse irq_desc.irq_data.affinity instead.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Ben Hutchings \u003cbhutchings@solarflare.com\u003e\n" }, { "commit": "cd7eab44e9946c28d595abe3e9a43e945bc49141", "tree": "b8b0c1beebbcad7785422a463504a374bb43cdea", "parents": [ "1bae4ce27c9c90344f23c65ea6966c50ffeae2f5" ], "author": { "name": "Ben Hutchings", "email": "bhutchings@solarflare.com", "time": "Wed Jan 19 21:01:44 2011 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Jan 22 17:34:25 2011 +0100" }, "message": "genirq: Add IRQ affinity notifiers\n\nWhen initiating I/O on a multiqueue and multi-IRQ device, we may want\nto select a queue for which the response will be handled on the same\nor a nearby CPU. This requires a reverse-map of IRQ affinity. Add a\nnotification mechanism to support this.\n\nThis is based closely on work by Thomas Gleixner \u003ctglx@linutronix.de\u003e.\n\nSigned-off-by: Ben Hutchings \u003cbhutchings@solarflare.com\u003e\nCc: linux-net-drivers@solarflare.com\nCc: Tom Herbert \u003ctherbert@google.com\u003e\nCc: David Miller \u003cdavem@davemloft.net\u003e\nLKML-Reference: \u003c1295470904.11126.84.camel@bwh-desktop\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "c9b5f501ef1580faa30c40c644b7691870462201", "tree": "d22df12b72bc90b130254078ee71b1b7e7e3e521", "parents": [ "524429c31b486c05449666b94613f59f729c0a84" ], "author": { "name": "Peter Zijlstra", "email": "a.p.zijlstra@chello.nl", "time": "Fri Jan 07 13:41:40 2011 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Jan 07 15:55:45 2011 +0100" }, "message": "sched: Constify function scope static struct sched_param usage\n\nFunction-scope statics are discouraged because they are\neasily overlooked and can cause subtle bugs/races due to\ntheir global (non-SMP safe) nature.\n\nLinus noticed that we did this for sched_param - at minimum\nmake the const.\n\nSuggested-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nLKML-Reference: Message-ID: \u003cAANLkTinotRxScOHEb0HgFgSpGPkq_6jKTv5CfvnQM\u003dee@mail.gmail.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "92fd4d4d67b945c0766416284d4ab236b31542c4", "tree": "00b8b5f90748f752ccaba3dddbe271091d93543a", "parents": [ "fe7de49f9d4e53f24ec9ef762a503f70b562341c", "e53beacd23d9cb47590da6a7a7f6d417b941a994" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Nov 18 13:22:14 2010 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Nov 18 13:22:26 2010 +0100" }, "message": "Merge commit \u0027v2.6.37-rc2\u0027 into sched/core\n\nMerge reason: Move to a .37-rc base.\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "fe7de49f9d4e53f24ec9ef762a503f70b562341c", "tree": "766b012c7199f12625bc909748175d9e37b0b87d", "parents": [ "d4429f608abde89e8bc1e24b43cd503feb95c496" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Oct 20 16:01:12 2010 -0700" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Sat Oct 23 17:56:48 2010 +0200" }, "message": "sched: Make sched_param argument static in sched_setscheduler() callers\n\nAndrew Morton pointed out almost all sched_setscheduler() callers are\nusing fixed parameters and can be converted to static. It reduces runtime\nmemory use a little.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReported-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nAcked-by: James Morris \u003cjmorris@namei.org\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nCc: Steven Rostedt \u003crostedt@goodmis.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "2656c36699677238edc9ec1fea79039f1fddbcb6", "tree": "4fd8b68709598b5c581b3792ad2fc7481be76e12", "parents": [ "d4429f608abde89e8bc1e24b43cd503feb95c496" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Oct 22 14:47:57 2010 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Oct 22 16:10:29 2010 +0200" }, "message": "genirq: Warn if enable_irq is called before irq is set up\n\nThe recent changes in the genirq core unearthed a bug in arch/um which\ncalled enable_irq() before the interrupt was set up.\n\nWarn and return instead of crashing the machine with a NULL pointer\ndereference.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Richard Weinberger \u003crichard@nod.at\u003e\n" }, { "commit": "2f7e99bb9be6a2d8d7b808dc86037710cc8b7bf1", "tree": "c620f5e8eb1a160bebb73a330fecd95a430977ed", "parents": [ "b2ba2c30033c10cca2454f8b44bf98f5249e61c6" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:45:50 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:48 2010 +0200" }, "message": "genirq: Provide compat handling for chip-\u003eset_wake()\n\nWrap the old chip function set_wake() until the migration is complete\nand the old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121842.927527393@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "b2ba2c30033c10cca2454f8b44bf98f5249e61c6", "tree": "a168cbcd1314f76edccecd5984d75793773b5239", "parents": [ "c96b3b3c448592a0b87ef20306deb8b1fb4878c7" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:45:47 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:47 2010 +0200" }, "message": "genirq: Provide compat handling for chip-\u003eset_type()\n\nWrap the old chip function set_type() until the migration is complete\nand the old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121842.832261548@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "c96b3b3c448592a0b87ef20306deb8b1fb4878c7", "tree": "744c6879a585b9ab07ed51315930f8a673c09aed", "parents": [ "37e12df709f09eac17314d79a52190ac46746e33" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:45:41 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:46 2010 +0200" }, "message": "genirq: Provide compat handling for chip-\u003eset_affinity()\n\nWrap the old chip function set_affinity() until the migration is\ncomplete and the old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121842.732894108@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "37e12df709f09eac17314d79a52190ac46746e33", "tree": "78159578101faddef22c5b13ccc200542cb3d2b6", "parents": [ "bc310dda41be6439364c8f3b9fe7c9d743d22b1c" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:45:38 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:44 2010 +0200" }, "message": "genirq: Provide compat handling for chip-\u003estartup()\n\nWrap the old chip function startup() until the migration is complete and\nthe old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121842.635152961@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "bc310dda41be6439364c8f3b9fe7c9d743d22b1c", "tree": "86a8b4da3d9e81aad104b61023858813454b0f35", "parents": [ "c5f756344c390f629243b4a28c2bd198fdfd7ee9" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:45:02 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:43 2010 +0200" }, "message": "genirq: Provide compat handling for chip-\u003edisable()/shutdown()\n\nWrap the old chip functions disable() and shutdown() until the\nmigration is complete and the old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121842.532070631@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "0eda58b7f3a30c9a13d83db1cfaab00e1c452055", "tree": "6e6efb689188f20271d8dd290008652d3c3defd8", "parents": [ "e2c0f8ff0fc26959952fbfa89f732fef928df77f" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:44:44 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:37 2010 +0200" }, "message": "genirq: Provide compat handling for chip-\u003eunmask()\n\nWrap the old chip function unmask() until the migration is complete\nand the old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121842.043608928@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "3876ec9ef3775d062345b3760d3271ecb8cd3fea", "tree": "6c1cd233753fe09d48bdf107da34b4dbe4e003fc", "parents": [ "a77c4635915021c646cc017f22239e66d1aab4d5" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Sep 27 12:44:35 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:43:35 2010 +0200" }, "message": "genirq: Provide compat handling for bus_lock/bus_sync_unlock\n\nWrap the old chip functions for bus_lock/bus_sync_unlock until the\nmigration is complete and the old chip functions are removed.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nLKML-Reference: \u003c20100927121841.842536121@linutronix.de\u003e\nReviewed-by: H. Peter Anvin \u003chpa@zytor.com\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "6b8ff3120c758340505dddf08ad685ebb841d5d5", "tree": "794eed27c6f9a8931b8fdf4a7ae60a1560b237fc", "parents": [ "ff7dcd44dd446db2c3e13bdedf2d52b8e0127f16" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Oct 01 12:58:38 2010 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Oct 04 12:36:26 2010 +0200" }, "message": "genirq: Convert core code to irq_data\n\nConvert all references in the core code to orq, chip, handler_data,\nchip_data, msi_desc, affinity to irq_data.*\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "685fd0b4ea3f0f1d5385610b0d5b57775a8d5842", "tree": "44c224195e6500c67812b38ff9fd60c96c6b8dc5", "parents": [ "1a041a23da7c77b53c71fe11b4f940388bee37b1" ], "author": { "name": "Ian Campbell", "email": "ian.campbell@citrix.com", "time": "Thu Jul 29 11:16:32 2010 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Jul 29 13:24:57 2010 +0200" }, "message": "irq: Add new IRQ flag IRQF_NO_SUSPEND\n\nA small number of users of IRQF_TIMER are using it for the implied no\nsuspend behaviour on interrupts which are not timer interrupts.\n\nTherefore add a new IRQF_NO_SUSPEND flag, rename IRQF_TIMER to\n__IRQF_TIMER and redefine IRQF_TIMER in terms of these new flags.\n\nSigned-off-by: Ian Campbell \u003cian.campbell@citrix.com\u003e\nCc: Jeremy Fitzhardinge \u003cjeremy@goop.org\u003e\nCc: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\nCc: Benjamin Herrenschmidt \u003cbenh@kernel.crashing.org\u003e\nCc: Paul Mackerras \u003cpaulus@samba.org\u003e\nCc: Grant Likely \u003cgrant.likely@secretlab.ca\u003e\nCc: xen-devel@lists.xensource.com\nCc: linux-input@vger.kernel.org\nCc: linuxppc-dev@ozlabs.org\nCc: devicetree-discuss@lists.ozlabs.org\nLKML-Reference: \u003c1280398595-29708-1-git-send-email-ian.campbell@citrix.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "4673247562e39a17e09440fa1400819522ccd446", "tree": "31746ff2cd7f41048b7507cd3cae9dae2a1519be", "parents": [ "84f7586edab93052bbe7eb72e1e9bce5f7b6ae45" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Jun 07 17:53:51 2010 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Jun 09 17:05:08 2010 +0200" }, "message": "genirq: Deal with desc-\u003eset_type() changing desc-\u003echip\n\nThe set_type() function can change the chip implementation when the\ntrigger mode changes. That might result in using an non-initialized\nirq chip when called from __setup_irq() or when called via\nset_irq_type() on an already enabled irq. \n\nThe set_irq_type() function should not be called on an enabled irq,\nbut because we forgot to put a check into it, we have a bunch of users\nwhich grew the habit of doing that and it never blew up as the\nfunction is serialized via desc-\u003elock against all users of desc-\u003echip\nand they never hit the non-initialized irq chip issue.\n\nThe easy fix for the __setup_irq() issue would be to move the\nirq_chip_set_defaults(desc-\u003echip) call after the trigger setting to\nmake sure that a chip change is covered.\n\nBut as we have already users, which do the type setting after\nrequest_irq(), the safe fix for now is to call irq_chip_set_defaults()\nfrom __irq_set_trigger() when desc-\u003eset_type() changed the irq chip.\n\nIt needs a deeper analysis whether we should refuse to change the chip\non an already enabled irq, but that\u0027d be a large scale change to fix\nall the existing users. So that\u0027s neither stable nor 2.6.35 material.\n\nReported-by: Esben Haabendal \u003ceha@doredevelopment.dk\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Benjamin Herrenschmidt \u003cbenh@kernel.crashing.org\u003e\nCc: linuxppc-dev \u003clinuxppc-dev@ozlabs.org\u003e\nCc: stable@kernel.org\n" }, { "commit": "e7a297b0d7d6049bd4e423ac1e17da31e4c401b8", "tree": "f10f12806a637b09bec89ef5428d981c4c1a4bc9", "parents": [ "6932bf37bed45ce8ed531928b1b0f98162fe6df6" ], "author": { "name": "Peter P Waskiewicz Jr", "email": "peter.p.waskiewicz.jr@intel.com", "time": "Fri Apr 30 14:44:50 2010 -0700" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon May 03 11:50:57 2010 +0200" }, "message": "genirq: Add CPU mask affinity hint\n\nThis patch adds a cpumask affinity hint to the irq_desc structure,\nalong with a registration function and a read-only proc entry for each\ninterrupt.\n\nThis affinity_hint handle for each interrupt can be used by underlying\ndrivers that need a better mechanism to control interrupt affinity.\nThe underlying driver can register a cpumask for the interrupt, which\nwill allow the driver to provide the CPU mask for the interrupt to\nanything that requests it. The intent is to extend the userspace\ndaemon, irqbalance, to help hint to it a preferred CPU mask to balance\nthe interrupt into.\n\n[ tglx: Fixed compile warnings, added WARN_ON, made SMP only ]\n\nSigned-off-by: Peter P Waskiewicz Jr \u003cpeter.p.waskiewicz.jr@intel.com\u003e\nCc: davem@davemloft.net\nCc: arjan@linux.jf.intel.com\nCc: bhutchings@solarflare.com\nLKML-Reference: \u003c20100430214445.3992.41647.stgit@ppwaskie-hc2.jf.intel.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "6932bf37bed45ce8ed531928b1b0f98162fe6df6", "tree": "3b7af6b49608cb8e3474bf6cb56f86e712466942", "parents": [ "e58aa3d2d0cc01ad8d6f7f640a0670433f794922" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Mar 26 00:06:55 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Apr 13 16:36:40 2010 +0200" }, "message": "genirq: Remove IRQF_DISABLED from core code\n\nRemove all code which is related to IRQF_DISABLED from the core kernel\ncode. IRQF_DISABLED still exists as a flag, but becomes a NOOP and\nwill be removed after a grace period. That way we can easily revert to\nthe previous behaviour by just restoring the core code.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Alan Cox \u003calan@lxorguk.ukuu.org.uk\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: David Miller \u003cdavem@davemloft.net\u003e\nCc: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\nCc: Arnaldo Carvalho de Melo \u003cacme@redhat.com\u003e\nCc: Linus Torvalds \u003ctorvalds@osdl.org\u003e\nLKML-Reference: \u003c20100326000405.991244690@linutronix.de\u003e\n" }, { "commit": "ae731f8d0785ccd3380f511bae888933b6562e45", "tree": "07db03ce79231153a4ae5df75c0ca4dcd96307c2", "parents": [ "7c7145f6acc68100dbdc5d3c5c64fe3af1c99c89" ], "author": { "name": "Marc Zyngier", "email": "maz@misterjones.org", "time": "Mon Mar 15 22:56:33 2010 +0000" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Apr 13 16:36:39 2010 +0200" }, "message": "genirq: Introduce request_any_context_irq()\n\nNow that we enjoy threaded interrupts, we\u0027re starting to see irq_chip\nimplementations (wm831x, pca953x) that make use of threaded interrupts\nfor the controller, and nested interrupts for the client interrupt. It\nall works very well, with one drawback:\n\nDrivers requesting an IRQ must now know whether the handler will\nrun in a thread context or not, and call request_threaded_irq() or\nrequest_irq() accordingly.\n\nThe problem is that the requesting driver sometimes doesn\u0027t know\nabout the nature of the interrupt, specially when the interrupt\ncontroller is a discrete chip (typically a GPIO expander connected\nover I2C) that can be connected to a wide variety of otherwise perfectly\nsupported hardware.\n\nThis patch introduces the request_any_context_irq() function that mostly\nmimics the usual request_irq(), except that it checks whether the irq\nlevel is configured as nested or not, and calls the right backend.\nOn success, it also returns either IRQC_IS_HARDIRQ or IRQC_IS_NESTED.\n\n[ tglx: Made return value an enum, simplified code and made the export\n \tof request_any_context_irq GPL ]\n\nSigned-off-by: Marc Zyngier \u003cmaz@misterjones.org\u003e\nCc: \u003cjoachim.eastwood@jotron.com\u003e\nLKML-Reference: \u003c927ea285bd0c68934ddae1a47e44a9ba@localhost\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "753649dbc49345a73a2454c770a3f2d54d11aec6", "tree": "7489a4f883a1281a4874817826ebf369f7d019f0", "parents": [ "2eaa9cfdf33b8d7fb7aff27792192e0019ae8fc6" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Mar 31 13:30:19 2010 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Mar 31 15:48:38 2010 +0200" }, "message": "genirq: Force MSI irq handlers to run with interrupts disabled\n\nNetwork folks reported that directing all MSI-X vectors of their multi\nqueue NICs to a single core can cause interrupt stack overflows when\nenough interrupts fire at the same time.\n\nThis is caused by the fact that we run interrupt handlers by default\nwith interrupts enabled unless the driver reuqests the interrupt with\nthe IRQF_DISABLED set. The NIC handlers do not set this flag, so\nsimultaneous interrupts can nest unlimited and cause the stack\noverflow.\n\nThe only safe counter measure is to run the interrupt handlers with\ninterrupts disabled. We can\u0027t switch to this mode in general right\nnow, but it is safe to do so for MSI interrupts.\n\nForce IRQF_DISABLED for MSI interrupt handlers.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Linus Torvalds \u003ctorvalds@osdl.org\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Alan Cox \u003calan@lxorguk.ukuu.org.uk\u003e\nCc: David Miller \u003cdavem@davemloft.net\u003e\nCc: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\nCc: Arnaldo Carvalho de Melo \u003cacme@redhat.com\u003e\nCc: stable@kernel.org\n" }, { "commit": "cc8c3b78433222e5dbc1fdfcfdde29e1743f181a", "tree": "7efe13d87d7d92b1ff409455e65c85129bc30121", "parents": [ "0b1adaa031a55e44f5dd942f234bf09d28e8a0d6" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 23 22:40:53 2010 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Mar 24 14:38:23 2010 +0100" }, "message": "genirq: Protect access to irq_desc-\u003eaction in can_request_irq()\n\ncan_request_irq() accesses and dereferences irq_desc-\u003eaction w/o\nholding irq_desc-\u003elock. So action can be freed on another CPU before\nit\u0027s dereferenced. Unlikely, but ...\n\nProtect it with desc-\u003elock.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "0b1adaa031a55e44f5dd942f234bf09d28e8a0d6", "tree": "354aa6cbfbcd856226c543b9f263f87864245065", "parents": [ "522dba7134d6b2e5821d3457f7941ec34f668e6d" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 09 19:45:54 2010 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Mar 10 17:45:14 2010 +0100" }, "message": "genirq: Prevent oneshot irq thread race\n\nLars-Peter pointed out that the oneshot threaded interrupt handler\ncode has the following race:\n\n CPU0 CPU1\n hande_level_irq(irq X)\n mask_ack_irq(irq X)\n handle_IRQ_event(irq X)\n wake_up(thread_handler)\n thread handler(irq X) runs\n finalize_oneshot(irq X)\n\t\t\t\t does not unmask due to \n\t\t\t\t !(desc-\u003estatus \u0026 IRQ_MASKED)\n\n return from irq\n does not unmask due to\n (desc-\u003estatus \u0026 IRQ_ONESHOT)\n \t\t\t\t \nThis leaves the interrupt line masked forever. \n\nThe reason for this is the inconsistent handling of the IRQ_MASKED\nflag. Instead of setting it in the mask function the oneshot support\nsets the flag after waking up the irq thread.\n\nThe solution for this is to set/clear the IRQ_MASKED status whenever\nwe mask/unmask an interrupt line. That\u0027s the easy part, but that\ncleanup opens another race:\n\n CPU0 CPU1\n hande_level_irq(irq)\n mask_ack_irq(irq)\n handle_IRQ_event(irq)\n wake_up(thread_handler)\n thread handler(irq) runs\n finalize_oneshot_irq(irq)\n\t\t\t\t unmask(irq)\n irq triggers again\n handle_level_irq(irq)\n mask_ack_irq(irq)\n return from irq due to IRQ_INPROGRESS\t\t\t\t \n\n return from irq\n does not unmask due to\n (desc-\u003estatus \u0026 IRQ_ONESHOT)\n\nThis requires that we synchronize finalize_oneshot_irq() with the\nprimary handler. If IRQ_INPROGESS is set we wait until the primary\nhandler on the other CPU has returned before unmasking the interrupt\nline again.\n\nWe probably have never seen that problem because it does not happen on\nUP and on SMP the irqbalancer protects us by pinning the primary\nhandler and the thread to the same CPU.\n\nReported-by: Lars-Peter Clausen \u003clars@metafoo.de\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: stable@kernel.org\n" }, { "commit": "239007b8440abff689632f50cdf0f2b9e895b534", "tree": "569cab843af4a999d6d868ec9a824530d2bfa733", "parents": [ "9f5a5621e78cf48d86682a71ceb3fcdbde38b222" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Nov 17 16:46:45 2009 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Dec 14 23:55:33 2009 +0100" }, "message": "genirq: Convert irq_desc.lock to raw_spinlock\n\nConvert locks which cannot be sleeping locks in preempt-rt to\nraw_spinlocks.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nAcked-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "2b876f95d03e226394b5d360c86127cbefaf614b", "tree": "1f7ea2b85896bac5f5caecb6ff7a8c5d5262914f", "parents": [ "fbf07eac7bf21c262143194181bd97c5d18b8ceb", "c95b4502ad7fe8f3b9954aec794b00ac0046ab3a", "6ce51c431019310ca03371355a4366c4649fa349" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Dec 08 19:30:19 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Dec 08 19:30:19 2009 -0800" }, "message": "Merge branches \u0027timers-for-linus-ntp\u0027 and \u0027irq-core-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip\n\n* \u0027timers-for-linus-ntp\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:\n ntp: Provide compability defines (You say MOD_NANO, I say ADJ_NANO)\n\n* \u0027irq-core-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:\n genirq: do not execute DEBUG_SHIRQ when irq setup failed\n" }, { "commit": "d90a7e86401ffea2163a4337f3a47f3909c4e255", "tree": "afe3d4aa24849c561bb1e6c04c4c54808cbedb47", "parents": [ "12a499612e1ff439bdad240c7f86c55366941d4d", "4dbc9ca219b0f294332e734528f7b82211700170" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Sep 11 13:21:31 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Sep 11 13:21:31 2009 -0700" }, "message": "Merge branch \u0027irq-threaded-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip\n\n* \u0027irq-threaded-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:\n genirq: Do not mask oneshot edge type interrupts\n genirq: Support nested threaded irq handling\n genirq: Add buslock support\n genirq: Add oneshot support\n" }, { "commit": "69ab849439b506cd8dd2879527fdb64d95dd5211", "tree": "059fba8bcb8147fafcd4e86db61f0f88d6a84095", "parents": [ "2d860ad76f4ee4d2eba0fe3797c8d7cdce432cc0" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Aug 17 14:07:16 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Aug 18 17:22:43 2009 +0200" }, "message": "genirq: Wake up irq thread after action has been installed\n\nThe wake_up_process() of the new irq thread in __setup_irq() is too\nearly as the irqaction is not yet fully initialized especially\naction-\u003eirq is not yet set. The interrupt thread might dereference the\nwrong irq descriptor.\n\nMove the wakeup after the action is installed and action-\u003eirq has been\nset.\n\nReported-by: Michael Buesch \u003cmb@bu3sch.de\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nTested-by: Michael Buesch \u003cmb@bu3sch.de\u003e\n" }, { "commit": "399b5da29b9f851eb7b96e2882097127f003e87c", "tree": "0264a7ae22988e8a298407c611b5b639c9315ff7", "parents": [ "70aedd24d20e75198f5a0b11750faabbb56924e2" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Aug 13 13:21:38 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Aug 17 10:54:05 2009 +0200" }, "message": "genirq: Support nested threaded irq handling\n\nInterrupt chips which are behind a slow bus (i2c, spi ...) and\ndemultiplex other interrupt sources need to run their interrupt\nhandler in a thread. \n\nThe demultiplexed interrupt handlers need to run in thread context as\nwell and need to finish before the demux handler thread can reenable\nthe interrupt line. So the easiest way is to run the sub device\nhandlers in the context of the demultiplexing handler thread.\n\nTo avoid that a separate thread is created for the subdevices the\nfunction set_nested_irq_thread() is provided which sets the\nIRQ_NESTED_THREAD flag in the interrupt descriptor.\n\nA driver which calls request_threaded_irq() must not be aware of the\nfact that the threaded handler is called in the context of the\ndemultiplexing handler thread. The setup code checks the\nIRQ_NESTED_THREAD flag which was set from the irq chip setup code and\ndoes not setup a separate thread for the interrupt. The primary\nfunction which is provided by the device driver is replaced by an\ninternal dummy function which warns when it is called.\n\nFor the demultiplexing handler a helper function handle_nested_irq()\nis provided which calls the demux interrupt thread function in the\ncontext of the caller and does the proper interrupt accounting and\ntakes the interrupt disabled status of the demultiplexed subdevice\ninto account.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Mark Brown \u003cbroonie@opensource.wolfsonmicro.com\u003e\nCc: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\nCc: Trilok Soni \u003csoni.trilok@gmail.com\u003e\nCc: Pavel Machek \u003cpavel@ucw.cz\u003e\nCc: Brian Swetland \u003cswetland@google.com\u003e\nCc: Joonyoung Shim \u003cjy0922.shim@samsung.com\u003e\nCc: m.szyprowski@samsung.com\nCc: t.fujak@samsung.com\nCc: kyungmin.park@samsung.com,\nCc: David Brownell \u003cdavid-b@pacbell.net\u003e\nCc: Daniel Ribeiro \u003cdrwyrm@gmail.com\u003e\nCc: arve@android.com\nCc: Barry Song \u003c21cnbao@gmail.com\u003e\n\n" }, { "commit": "70aedd24d20e75198f5a0b11750faabbb56924e2", "tree": "8492641c61aa3af6f4dea421b8f628efe6fc92bd", "parents": [ "b25c340c195447afb1860da580fe2a85a6b652c5" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Aug 13 12:17:48 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Aug 17 10:54:05 2009 +0200" }, "message": "genirq: Add buslock support\n\nSome interrupt chips are connected to a \"slow\" bus (i2c, spi ...). The\nbus access needs to sleep and therefor cannot be called in atomic\ncontexts.\n\nSome of the generic interrupt management functions like disable_irq(),\nenable_irq() ... call interrupt chip functions with the irq_desc-\u003elock\nheld and interrupts disabled. This does not work for such devices.\n\nProvide a separate synchronization mechanism for such interrupt\nchips. The irq_chip structure is extended by two optional functions\n(bus_lock and bus_sync_and_unlock).\n\nThe idea is to serialize the bus access for those operations in the\ncore code so that drivers which are behind that bus operated interrupt\ncontroller do not have to worry about it and just can use the normal\ninterfaces. To achieve this we add two function pointers to the\nirq_chip: bus_lock and bus_sync_unlock.\n\nbus_lock() is called to serialize access to the interrupt controller\nbus.\n\nNow the core code can issue chip-\u003emask/unmask ... commands without\nchanging the fast path code at all. The chip implementation merily\nstores that information in a chip private data structure and\nreturns. No bus interaction as these functions are called from atomic\ncontext.\n\nAfter that bus_sync_unlock() is called outside the atomic context. Now\nthe chip implementation issues the bus commands, waits for completion\nand unlocks the interrupt controller bus.\n\nThe irq_chip implementation as pseudo code:\n\nstruct irq_chip_data {\n struct mutex mutex;\n unsigned int irq_offset;\n unsigned long mask;\n unsigned long mask_status;\n}\n\nstatic void bus_lock(unsigned int irq)\n{\n struct irq_chip_data *data \u003d get_irq_desc_chip_data(irq);\n\n mutex_lock(\u0026data-\u003emutex);\n}\n\nstatic void mask(unsigned int irq)\n{\n struct irq_chip_data *data \u003d get_irq_desc_chip_data(irq);\n\n irq -\u003d data-\u003eirq_offset;\n data-\u003emask |\u003d (1 \u003c\u003c irq);\n}\n\nstatic void unmask(unsigned int irq)\n{\n struct irq_chip_data *data \u003d get_irq_desc_chip_data(irq);\n\n irq -\u003d data-\u003eirq_offset;\n data-\u003emask \u0026\u003d ~(1 \u003c\u003c irq);\n}\n\nstatic void bus_sync_unlock(unsigned int irq)\n{\n struct irq_chip_data *data \u003d get_irq_desc_chip_data(irq);\n\n if (data-\u003emask !\u003d data-\u003emask_status) {\n do_bus_magic_to_set_mask(data-\u003emask);\n data-\u003emask_status \u003d data-\u003emask;\n }\n mutex_unlock(\u0026data-\u003emutex);\n}\n\nThe device drivers can use request_threaded_irq, free_irq, disable_irq\nand enable_irq as usual with the only restriction that the calls need\nto come from non atomic context.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Mark Brown \u003cbroonie@opensource.wolfsonmicro.com\u003e\nCc: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\nCc: Trilok Soni \u003csoni.trilok@gmail.com\u003e\nCc: Pavel Machek \u003cpavel@ucw.cz\u003e\nCc: Brian Swetland \u003cswetland@google.com\u003e\nCc: Joonyoung Shim \u003cjy0922.shim@samsung.com\u003e\nCc: m.szyprowski@samsung.com\nCc: t.fujak@samsung.com\nCc: kyungmin.park@samsung.com,\nCc: David Brownell \u003cdavid-b@pacbell.net\u003e\nCc: Daniel Ribeiro \u003cdrwyrm@gmail.com\u003e\nCc: arve@android.com\nCc: Barry Song \u003c21cnbao@gmail.com\u003e\n\n" }, { "commit": "b25c340c195447afb1860da580fe2a85a6b652c5", "tree": "99578092c1e6100f50786e44a1a4fe9cfdb11662", "parents": [ "b2add73dbf93fd50f00564d7abc3e2b9aa9dd20c" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Aug 13 12:17:22 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Aug 17 10:54:05 2009 +0200" }, "message": "genirq: Add oneshot support\n\nFor threaded interrupt handlers we expect the hard interrupt handler\npart to mask the interrupt on the originating device. The interrupt\nline itself is reenabled after the hard interrupt handler has\nexecuted.\n\nThis requires access to the originating device from hard interrupt\ncontext which is not always possible. There are devices which can only\nbe accessed via a bus (i2c, spi, ...). The bus access requires thread\ncontext. For such devices we need to keep the interrupt line masked\nuntil the threaded handler has executed.\n\nAdd a new flag IRQF_ONESHOT which allows drivers to request that the\ninterrupt is not unmasked after the hard interrupt context handler has\nbeen executed and the thread has been woken. The interrupt line is\nunmasked after the thread handler function has been executed.\n\nNote that for now IRQF_ONESHOT cannot be used with IRQF_SHARED to\navoid complex accounting mechanisms.\n\nFor oneshot interrupts the primary handler simply returns\nIRQ_WAKE_THREAD and does nothing else. A generic implementation\nirq_default_primary_handler() is provided to avoid useless copies all\nover the place. It is automatically installed when\nrequest_threaded_irq() is called with handler\u003dNULL and\nthread_fn!\u003dNULL.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Mark Brown \u003cbroonie@opensource.wolfsonmicro.com\u003e\nCc: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\nCc: Trilok Soni \u003csoni.trilok@gmail.com\u003e\nCc: Pavel Machek \u003cpavel@ucw.cz\u003e\nCc: Brian Swetland \u003cswetland@google.com\u003e\nCc: Joonyoung Shim \u003cjy0922.shim@samsung.com\u003e\nCc: m.szyprowski@samsung.com\nCc: t.fujak@samsung.com\nCc: kyungmin.park@samsung.com,\nCc: David Brownell \u003cdavid-b@pacbell.net\u003e\nCc: Daniel Ribeiro \u003cdrwyrm@gmail.com\u003e\nCc: arve@android.com\nCc: Barry Song \u003c21cnbao@gmail.com\u003e\n\n" }, { "commit": "2d860ad76f4ee4d2eba0fe3797c8d7cdce432cc0", "tree": "928e3a162ce9bf9294afd138f5609cf3b70a747b", "parents": [ "3493e84de60590d3012139187f631f2dfbf0887f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 13 13:05:10 2009 -0700" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Aug 13 23:09:27 2009 +0200" }, "message": "genirq: prevent wakeup of freed irq thread\n\nfree_irq() can remove an irqaction while the corresponding interrupt\nis in progress, but free_irq() sets action-\u003ethread to NULL\nunconditionally, which might lead to a NULL pointer dereference in\nhandle_IRQ_event() when the hard interrupt context tries to wake up\nthe handler thread.\n\nPrevent this by moving the thread stop after synchronize_irq(). No\nneed to set action-\u003ethread to NULL either as action is going to be\nfreed anyway.\n\nThis fixes a boot crash reported against preempt-rt which uses the\nmainline irq threads code to implement full irq threading.\n\n[ tglx: removed local irqthread variable ]\n\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "61f3826133dc07142935fb5712fc738e19eb5575", "tree": "2bac9166734e564a4f89d9db9d7498df87597c31", "parents": [ "3c3301083e3bea4d14c597106c7b20b4b85fc03d" ], "author": { "name": "Bruno Premont", "email": "bonbons@linux-vserver.org", "time": "Wed Jul 22 22:22:32 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Wed Jul 22 23:18:46 2009 +0200" }, "message": "genirq: Fix UP compile failure caused by irq_thread_check_affinity\n\nSince genirq: Delegate irq affinity setting to the irq thread\n(591d2fb02ea80472d846c0b8507007806bdd69cc) compilation with\nCONFIG_SMP\u003dn fails with following error:\n\n/usr/src/linux-2.6/kernel/irq/manage.c:\n In function \u0027irq_thread_check_affinity\u0027:\n/usr/src/linux-2.6/kernel/irq/manage.c:475:\n error: \u0027struct irq_desc\u0027 has no member named \u0027affinity\u0027\nmake[4]: *** [kernel/irq/manage.o] Error 1\n\nThat commit adds a new function irq_thread_check_affinity() which\nuses struct irq_desc.affinity which is only available for CONFIG_SMP\u003dy.\nMove that function under #ifdef CONFIG_SMP.\n\n[ tglx@brownpaperbag: compile and boot tested on UP and SMP ]\n\nSigned-off-by: Bruno Premont \u003cbonbons@linux-vserver.org\u003e\nLKML-Reference: \u003c20090722222232.2eb3e1c4@neptune.home\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "591d2fb02ea80472d846c0b8507007806bdd69cc", "tree": "c7962a95a47bbdf664f15a504eff24c351f33613", "parents": [ "aea1f7964ae6cba5eb419a958956deb9016b3341" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Jul 21 11:09:39 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Jul 21 14:35:07 2009 +0200" }, "message": "genirq: Delegate irq affinity setting to the irq thread\n\nirq_set_thread_affinity() calls set_cpus_allowed_ptr() which might\nsleep, but irq_set_thread_affinity() is called with desc-\u003elock held\nand can be called from hard interrupt context as well. The code has\nanother bug as it does not hold a ref on the task struct as required\nby set_cpus_allowed_ptr().\n\nJust set the IRQTF_AFFINITY bit in action-\u003ethread_flags. The next time\nthe thread runs it migrates itself. Solves all of the above problems\nnicely.\n\nAdd kerneldoc to irq_set_thread_affinity() while at it.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nLKML-Reference: \u003cnew-submission\u003e\n\n" }, { "commit": "2453d6ff6ffc5f0d496b7b14f509a26f99bf115e", "tree": "9e4e709bdb932889b0b9cfc5b2b767cfcf750b83", "parents": [ "12e24f34cb0d55efd08c18b2112507d4bf498008", "ab33dcff40d7a9a28587e4425621e4cbc4089e03" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jun 20 11:30:01 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jun 20 11:30:01 2009 -0700" }, "message": "Merge branch \u0027irq-fixes-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip\n\n* \u0027irq-fixes-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:\n genirq, irq.h: Fix kernel-doc warnings\n genirq: fix comment to say IRQ_WAKE_THREAD\n" }, { "commit": "39a2eddb9b62959dc55c6978b5eaeb3dd57c5ff2", "tree": "cef63cdc205079cbdbeafd71998eacb31052b54c", "parents": [ "1d80cac0fe44fb87b2a3d35fddd7f534ea81cd90" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Tue May 12 14:35:54 2009 -0400" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue May 12 22:11:29 2009 +0200" }, "message": "genirq: fix comment to say IRQ_WAKE_THREAD\n\nTrying to implement a driver to use threaded irqs, I was confused when the \nreturn value to use that was described in the comment above \nrequest_threaded_irq was not defined.\n\nTurns out that the enum is IRQ_WAKE_THREAD where as the comment said \nIRQ_THREAD_WAKE.\n\n[Impact: do not confuse developers with wrong comments ]\n\nSigned-off-by: Steven Rostedt \u003crostedt@goodmis.org\u003e\nLKML-Reference: \u003calpine.DEB.2.00.0905121431020.13338@gandalf.stny.rr.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n\n" }, { "commit": "57b150cce8e004ddd36330490a68bfb59b7271e9", "tree": "f337b08b6cd9bbe165c709a7e5159b0d6d779238", "parents": [ "d5dedd4507d307eb3f35f21b6e16f336fdc0d82a" ], "author": { "name": "Yinghai Lu", "email": "yinghai@kernel.org", "time": "Mon Apr 27 17:59:53 2009 -0700" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Tue Apr 28 12:21:16 2009 +0200" }, "message": "irq: only update affinity if -\u003eset_affinity() is sucessfull\n\nirq_set_affinity() and move_masked_irq() try to assign affinity\nbefore calling chip set_affinity(). Some archs are assigning it\nin -\u003eset_affinity() again.\n\nWe do something like:\n\n cpumask_cpy(desc-\u003eaffinity, mask);\n desc-\u003echip-\u003eset_affinity(mask);\n\nBut in the failure path, affinity should not be touched - otherwise\nwe\u0027ll end up with a different affinity mask despite the failure to\nmigrate the IRQ.\n\nSo try to update the afffinity only if set_affinity returns with 0.\nAlso call irq_set_thread_affinity accordingly.\n\nv2: update after \"irq, x86: Remove IRQ_DISABLED check in process context IRQ move\"\nv3: according to Ingo, change set_affinity() in irq_chip should return int.\nv4: update comments by removing moving irq_desc code.\n\n[ Impact: fix /proc/irq/*/smp_affinity setting corner case bug ]\n\nSigned-off-by: Yinghai Lu \u003cyinghai@kernel.org\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nCc: Suresh Siddha \u003csuresh.b.siddha@intel.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nLKML-Reference: \u003c49F65509.60307@kernel.org\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "6ce51c431019310ca03371355a4366c4649fa349", "tree": "8f081f9c6ce501833f105e3d1541f748d9614adc", "parents": [ "833bb3046b6cb320e775ea2160ddca87d53260d5" ], "author": { "name": "Luis Henriques", "email": "henrix@sapo.pt", "time": "Wed Apr 01 18:06:35 2009 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Apr 23 08:45:48 2009 +0200" }, "message": "genirq: do not execute DEBUG_SHIRQ when irq setup failed\n\nWhen requesting an IRQ, the DEBUG_SHIRQ code executes a fake IRQ just to make\nsure the driver is ready to receive an IRQ immediately. The problem was that\nthis fake IRQ was being executed even if interrupt line failed to be allocated\nby __setup_irq.\n\nSigned-off-by: Luis Henriques \u003chenrix@sapo.pt\u003e\nLKML-Reference: \u003c20090401170635.GA4392@hades.domain.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n[ fixed bug pointed out by a warning reported by Stephen Rothwell ]\nCc: Stephen Rothwell \u003csfr@canb.auug.org.au\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "6ec3cfeca04622e3d80c9270191cd7f5f88214af", "tree": "ba9d66ec365e59842fe97d72855c6065763dfe98", "parents": [ "b21597d0268983f8f9e8b563494f75490403e948" ], "author": { "name": "Pallipadi, Venkatesh", "email": "venkatesh.pallipadi@intel.com", "time": "Mon Apr 13 15:20:58 2009 -0700" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Tue Apr 14 15:21:13 2009 +0200" }, "message": "x86, irq: Remove IRQ_DISABLED check in process context IRQ move\n\nAs discussed in the thread here:\n\n http://marc.info/?l\u003dlinux-kernel\u0026m\u003d123964468521142\u0026w\u003d2\n\nEric W. Biederman observed:\n\n\u003e It looks like some additional bugs have slipped in since last I looked.\n\u003e\n\u003e set_irq_affinity does this:\n\u003e ifdef CONFIG_GENERIC_PENDING_IRQ\n\u003e if (desc-\u003estatus \u0026 IRQ_MOVE_PCNTXT || desc-\u003estatus \u0026 IRQ_DISABLED) {\n\u003e cpumask_copy(desc-\u003eaffinity, cpumask);\n\u003e desc-\u003echip-\u003eset_affinity(irq, cpumask);\n\u003e } else {\n\u003e desc-\u003estatus |\u003d IRQ_MOVE_PENDING;\n\u003e cpumask_copy(desc-\u003epending_mask, cpumask);\n\u003e }\n\u003e #else\n\u003e\n\u003e That IRQ_DISABLED case is a software state and as such it has nothing to\n\u003e do with how safe it is to move an irq in process context.\n\n[...]\n\n\u003e\n\u003e The only reason we migrate MSIs in interrupt context today is that there\n\u003e wasn\u0027t infrastructure for support migration both in interrupt context\n\u003e and outside of it.\n\nYes. The idea here was to force the MSI migration to happen in process\ncontext. One of the patches in the series did\n\n disable_irq(dev-\u003eirq);\n irq_set_affinity(dev-\u003eirq, cpumask_of(dev-\u003ecpu));\n enable_irq(dev-\u003eirq);\n\nwith the above patch adding irq/manage code check for interrupt disabled\nand moving the interrupt in process context.\n\nIIRC, there was no IRQ_MOVE_PCNTXT when we were developing this HPET\ncode and we ended up having this ugly hack. IRQ_MOVE_PCNTXT was there\nwhen we eventually submitted the patch upstream. But, looks like I did a\nblind rebasing instead of using IRQ_MOVE_PCNTXT in hpet MSI code.\n\nBelow patch fixes this. i.e., revert commit 932775a4ab622e3c99bd59f14cc\nand add PCNTXT to HPET MSI setup. Also removes copying of desc-\u003eaffinity\nin generic code as set_affinity routines are doing it internally.\n\nReported-by: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nSigned-off-by: Venkatesh Pallipadi \u003cvenkatesh.pallipadi@intel.com\u003e\nAcked-by: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: \"Li Shaohua\" \u003cshaohua.li@intel.com\u003e\nCc: Gary Hade \u003cgaryhade@us.ibm.com\u003e\nCc: \"lcm@us.ibm.com\" \u003clcm@us.ibm.com\u003e\nCc: suresh.b.siddha@intel.com\nLKML-Reference: \u003c20090413222058.GB8211@linux-os.sc.intel.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "9efe21cb82b5dbe3b0b2ae4de4eccc64ecb94e95", "tree": "7ff8833745d2f268f897f6fa4a27263b4a572245", "parents": [ "de18836e447c2dc30120c0919b8db8ddc0401cc4", "0221c81b1b8eb0cbb6b30a0ced52ead32d2b4e4c" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Apr 06 01:41:22 2009 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Apr 06 01:41:22 2009 +0200" }, "message": "Merge branch \u0027linus\u0027 into irq/threaded\n\nConflicts:\n\tinclude/linux/irq.h\n\tkernel/irq/handle.c\n" }, { "commit": "0a0c5168df270a50e3518e4f12bddb31f8f5f38f", "tree": "ea1e392fef9d52a7b81c9010580c09317a4d707d", "parents": [ "019abbc87025a030fd25008612afd4eff8a375f7" ], "author": { "name": "Rafael J. Wysocki", "email": "rjw@sisk.pl", "time": "Mon Mar 16 22:33:49 2009 +0100" }, "committer": { "name": "Rafael J. Wysocki", "email": "rjw@sisk.pl", "time": "Mon Mar 30 21:46:54 2009 +0200" }, "message": "PM: Introduce functions for suspending and resuming device interrupts\n\nIntroduce helper functions allowing us to prevent device drivers from\ngetting any interrupts (without disabling interrupts on the CPU)\nduring suspend (or hibernation) and to make them start to receive\ninterrupts again during the subsequent resume. These functions make it\npossible to keep timer interrupts enabled while the \"late\" suspend and\n\"early\" resume callbacks provided by device drivers are being\nexecuted. In turn, this allows device drivers\u0027 \"late\" suspend and\n\"early\" resume callbacks to sleep, execute ACPI callbacks etc.\n\nThe functions introduced here will be used to rework the handling of\ninterrupts during suspend (hibernation) and resume. Namely,\ninterrupts will only be disabled on the CPU right before suspending\nsysdevs, while device drivers will be prevented from receiving\ninterrupts, with the help of the new helper function, before their\n\"late\" suspend callbacks run (and analogously during resume).\n\nSigned-off-by: Rafael J. Wysocki \u003crjw@sisk.pl\u003e\nAcked-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "6e15cf04860074ad032e88c306bea656bbdd0f22", "tree": "c346383bb7563e8d66b2f4a502f875b259c34870", "parents": [ "be0ea69674ed95e1e98cb3687a241badc756d228", "60db56422043aaa455ac7f858ce23c273220f9d9" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Mar 26 21:39:17 2009 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Mar 27 17:28:43 2009 +0100" }, "message": "Merge branch \u0027core/percpu\u0027 into percpu-cpumask-x86-for-linus-2\n\nConflicts:\n\tarch/parisc/kernel/irq.c\n\tarch/x86/include/asm/fixmap_64.h\n\tarch/x86/include/asm/setup.h\n\tkernel/irq/handle.c\n\nSemantic merge:\n arch/x86/include/asm/fixmap.h\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "f48fe81e5b032914183e9a17052313720c2cac56", "tree": "9404047368387a6f8b39b4a822ef28e18eba45e9", "parents": [ "935bd5b971f0df7c06d214d022cf8392e2f37952" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 24 11:46:22 2009 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 24 12:15:23 2009 +0100" }, "message": "genirq: threaded irq handlers review fixups\n\nDelta patch to address the review comments.\n\n - Implement warning when IRQ_WAKE_THREAD is requested and no\n thread handler installed\n - coding style fixes\n\nPointed-out-by: Christoph Hellwig \u003chch@infradead.org\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n\n" }, { "commit": "3aa551c9b4c40018f0e261a178e3d25478dc04a9", "tree": "2a696109273fcc421d774cc8fefa4180331a85ad", "parents": [ "80c5520811d3805adcb15c570ea5e2d489fa5d0b" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 23 18:28:15 2009 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Tue Mar 24 12:15:23 2009 +0100" }, "message": "genirq: add threaded interrupt handler support\n\nAdd support for threaded interrupt handlers:\n\nA device driver can request that its main interrupt handler runs in a\nthread. To achive this the device driver requests the interrupt with\nrequest_threaded_irq() and provides additionally to the handler a\nthread function. The handler function is called in hard interrupt\ncontext and needs to check whether the interrupt originated from the\ndevice. If the interrupt originated from the device then the handler\ncan either return IRQ_HANDLED or IRQ_WAKE_THREAD. IRQ_HANDLED is\nreturned when no further action is required. IRQ_WAKE_THREAD causes\nthe genirq code to invoke the threaded (main) handler. When\nIRQ_WAKE_THREAD is returned handler must have disabled the interrupt\non the device level. This is mandatory for shared interrupt handlers,\nbut we need to do it as well for obscure x86 hardware where disabling\nan interrupt on the IO_APIC level redirects the interrupt to the\nlegacy PIC interrupt lines.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nReviewed-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n\n" }, { "commit": "80c5520811d3805adcb15c570ea5e2d489fa5d0b", "tree": "ae797a7f4af39f80e77526533d06ac23b439f0ab", "parents": [ "b3e3b302cf6dc8d60b67f0e84d1fa5648889c038", "8c083f081d0014057901c68a0a3e0f8ca7ac8d23" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 23 14:50:03 2009 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 23 21:20:20 2009 +0100" }, "message": "Merge branch \u0027cpus4096\u0027 into irq/threaded\n\nConflicts:\n\tarch/parisc/kernel/irq.c\n\tkernel/irq/handle.c\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "edb35028e40515beb2f94808aff8b3e71fb9f35a", "tree": "153da9f8ca1934b30474bb15478feb14a7bd5b75", "parents": [ "8f8573ae9f5deefada6f5d64d0a52c9b39c730c7", "cb065c06b6cc615a58860d619d7fa7952cd6a18b", "5bee17f18b595937e6beafeee5197868a3f74a06" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Mar 16 09:20:13 2009 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Mar 16 09:20:13 2009 +0100" }, "message": "Merge branches \u0027irq/genirq\u0027 and \u0027linus\u0027 into irq/core\n" }, { "commit": "4553573277906901f62f73c0432b332c53de5e2c", "tree": "4093af83302fa711e9079152089085927f60c838", "parents": [ "bedd30d986a05e32dc3eab874e4b9ed8a38058bb" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sun Feb 22 23:00:32 2009 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Mar 13 14:32:29 2009 +0100" }, "message": "genirq: use kzalloc instead of explicit zero initialization\n\nImpact: simplification\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nReviewed-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\n" }, { "commit": "c8e2aeef0b8ac9fb8821b8b3734c031579d0b77a", "tree": "6b171eb497bdd5f21f9e6a0b06c4aa6c3c60b76b", "parents": [ "3dd3d46b78c22503957230ca5981849b7bb29b9a" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Mon Mar 09 20:26:23 2009 +0100" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Mar 13 14:32:28 2009 +0100" }, "message": "genirq: remove redundant if condition\n\nImpact: cleanup\n\nThe code is only compiled if CONFIG_GENERIC_HARDIRQS\u003dy so another\ncheck for this define in the code is redundant. Remove it.\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "eb53b4e8fef10ccccb49a6dbb5e19ca84ba5a305", "tree": "156b1a9b799c0b1c4dc90b05a7b2468df56ca1c1", "parents": [ "cbf94f06824780183e4bba165c7c29d5c7bd9a51" ], "author": { "name": "Magnus Damm", "email": "damm@igel.co.jp", "time": "Thu Mar 12 21:05:59 2009 +0900" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Mar 12 13:16:33 2009 +0100" }, "message": "irq: export remove_irq() and setup_irq() symbols\n\nExport the setup_irq() and remove_irq() symbols.\n\nI\u0027d like to export these functions since I have timer\ncode that needs to use setup_irq() early on (too early\nfor request_irq()), and the same code can also be\ncompiled as a module.\n\nSigned-off-by: Magnus Damm \u003cdamm@igel.co.jp\u003e\nLKML-Reference: \u003c20090312120559.2926.82371.sendpatchset@rx1.opensource.se\u003e\n[ changed to _GPL as these are special APIs deep inside the irq layer. ]\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" } ], "next": "cbf94f06824780183e4bba165c7c29d5c7bd9a51" }