)]}' { "log": [ { "commit": "0cdf6990e992902ae59cbc625d28cb41390f378e", "tree": "0c01cf792be5f36ea34064036005f424ab95a571", "parents": [ "de081fa517fed81b0369f2e90ca87c30182879c8", "cec7c893d8654723028f09d33341e42673558057" ], "author": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Jul 12 16:45:40 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Jul 12 16:45:40 2007 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (76 commits)\n IB: Update MAINTAINERS with Hal\u0027s new email address\n IB/mlx4: Implement query SRQ\n IB/mlx4: Implement query QP\n IB/cm: Send no match if a SIDR REQ does not match a listen\n IB/cm: Fix handling of duplicate SIDR REQs\n IB/cm: cm_msgs.h should include ib_cm.h\n IB/cm: Include HCA ACK delay in local ACK timeout\n IB/cm: Use spin_lock_irq() instead of spin_lock_irqsave() when possible\n IB/sa: Make sure SA queries use default P_Key\n IPoIB: Recycle loopback skbs instead of freeing and reallocating\n IB/mthca: Replace memset(\u003caddr\u003e, 0, PAGE_SIZE) with clear_page(\u003caddr\u003e)\n IPoIB/cm: Fix warning if IPV6 is not enabled\n IB/core: Take sizeof the correct pointer when calling kmalloc()\n IB/ehca: Improve latency by unlocking after triggering the hardware\n IB/ehca: Notify consumers of LID/PKEY/SM changes after nondisruptive events\n IB/ehca: Return QP pointer in poll_cq()\n IB/ehca: Change idr spinlocks into rwlocks\n IB/ehca: Refactor sync between completions and destroy_cq using atomic_t\n IB/ehca: Lock renaming, static initializers\n IB/ehca: Report RDMA atomic attributes in query_qp()\n ...\n" }, { "commit": "65541cb7cf353946ecd78016a453b453b8830656", "tree": "0a93ed27fdb6caf3f90250eaaf179b560e6742f3", "parents": [ "6a775e2ba4f7635849ade628e64723ab2beef0bc" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Jun 21 13:03:11 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jul 12 15:41:24 2007 -0700" }, "message": "IB/mlx4: Implement query SRQ\n\nSigned-off-by: Dotan Barak \u003cdotanb@mellanox.co.il\u003e\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "6a775e2ba4f7635849ade628e64723ab2beef0bc", "tree": "3c80d06d5f2f7135c81dddc2dcd6e536f9ecf22d", "parents": [ "6164c8cd1333403a28202f7c7e64ff9086d8f1aa" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Jun 21 12:27:47 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jul 12 15:41:00 2007 -0700" }, "message": "IB/mlx4: Implement query QP\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "21ba0f88ae56da82a3a15fe54d729208b64c4f4b", "tree": "17ce67f276fe3ea7284c3dc730bdd6a2ec7dfe2f", "parents": [ "dc690d8ef842b464f1c429a376ca16cb8dbee6ae", "36e235901f90fb83215be43cbd8f1ca14661ea40" ], "author": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Jul 12 13:40:57 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Jul 12 13:40:57 2007 -0700" }, "message": "Merge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6\n\n* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6: (34 commits)\n PCI: Only build PCI syscalls on architectures that want them\n PCI: limit pci_get_bus_and_slot to domain 0\n PCI: hotplug: acpiphp: avoid acpiphp \"cannot get bridge info\" PCI hotplug failure\n PCI: hotplug: acpiphp: remove hot plug parameter write to PCI host bridge\n PCI: hotplug: acpiphp: fix slot poweroff problem on systems without _PS3\n PCI: hotplug: pciehp: wait for 1 second after power off slot\n PCI: pci_set_power_state(): check for PM capabilities earlier\n PCI: cpci_hotplug: Convert to use the kthread API\n PCI: add pci_try_set_mwi\n PCI: pcie: remove SPIN_LOCK_UNLOCKED\n PCI: ROUND_UP macro cleanup in drivers/pci\n PCI: remove pci_dac_dma_... APIs\n PCI: pci-x-pci-express-read-control-interfaces cleanups\n PCI: Fix typo in include/linux/pci.h\n PCI: pci_ids, remove double or more empty lines\n PCI: pci_ids, add atheros and 3com_2 vendors\n PCI: pci_ids, reorder some entries\n PCI: i386: traps, change VENDOR to DEVICE\n PCI: ATM: lanai, change VENDOR to DEVICE\n PCI: Change all drivers to use pci_device-\u003erevision\n ...\n" }, { "commit": "7b595756ec1f49e0049a9e01a1298d53a7faaa15", "tree": "cd06687ab3e5c7a5a4ef91903dff207a18c4db76", "parents": [ "dbde0fcf9f8f6d477af3c32d9979e789ee680cde" ], "author": { "name": "Tejun Heo", "email": "htejun@gmail.com", "time": "Thu Jun 14 03:45:17 2007 +0900" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:09:06 2007 -0700" }, "message": "sysfs: kill unnecessary attribute-\u003eowner\n\nsysfs is now completely out of driver/module lifetime game. After\ndeletion, a sysfs node doesn\u0027t access anything outside sysfs proper,\nso there\u0027s no reason to hold onto the attribute owners. Note that\noften the wrong modules were accounted for as owners leading to\naccessing removed modules.\n\nThis patch kills now unnecessary attribute-\u003eowner. Note that with\nthis change, userland holding a sysfs node does not prevent the\nbacking module from being unloaded.\n\nFor more info regarding lifetime rule cleanup, please read the\nfollowing message.\n\n http://article.gmane.org/gmane.linux.kernel/510293\n\n(tweaked by Greg to not delete the field just yet, to make it easier to\nmerge things properly.)\n\nSigned-off-by: Tejun Heo \u003chtejun@gmail.com\u003e\nCc: Cornelia Huck \u003ccornelia.huck@de.ibm.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "44c10138fd4bbc4b6d6bff0873c24902f2a9da65", "tree": "6e16d3ec80c87490dc743f72da086356f2906ace", "parents": [ "b8a3a5214d7cc115f1ca3a3967b7229d97c46f4a" ], "author": { "name": "Auke Kok", "email": "auke-jan.h.kok@intel.com", "time": "Fri Jun 08 15:46:36 2007 -0700" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Wed Jul 11 16:02:10 2007 -0700" }, "message": "PCI: Change all drivers to use pci_device-\u003erevision\n\nInstead of all drivers reading pci config space to get the revision\nID, they can now use the pci_device-\u003erevision member.\n\nThis exposes some issues where drivers where reading a word or a dword\nfor the revision number, and adding useless error-handling around the\nread. Some drivers even just read it for no purpose of all.\n\nIn devices where the revision ID is being copied over and used in what\nappears to be the equivalent of hotpath, I have left the copy code\nand the cached copy as not to influence the driver\u0027s performance.\n\nCompile tested with make all{yes,mod}config on x86_64 and i386.\n\nSigned-off-by: Auke Kok \u003cauke-jan.h.kok@intel.com\u003e\nAcked-by: Dave Jones \u003cdavej@redhat.com\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "6164c8cd1333403a28202f7c7e64ff9086d8f1aa", "tree": "13f7156734c73a7f456572b252b8a8de87a65217", "parents": [ "29c2731cbfa9d5b560a73461775f2be03b46dd4c" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Mon Jul 02 16:14:14 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 21:52:28 2007 -0700" }, "message": "IB/cm: Send no match if a SIDR REQ does not match a listen\n\nIf a SIDR REQ does not match a listen, we should reply with status\nvalue 1 (service ID not supported), rather than dropping through to\nthe default case of status 2 (rejected by service provider).\n\nDoing this also fixes a bug where the cm_id_priv is removed from the\nremote_sidr_table twice.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "29c2731cbfa9d5b560a73461775f2be03b46dd4c", "tree": "c2f4edc90572fe3e782f54f4fc0c43171f657668", "parents": [ "5d861be8c87b8a8f8e0f58b3b8864f31da7ee9c3" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Mon Jul 02 14:36:19 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 21:51:43 2007 -0700" }, "message": "IB/cm: Fix handling of duplicate SIDR REQs\n\nFix handling to duplicate SIDR REQs to avoid sending a reject if a\nduplicate is detected. Duplicates should just be silently discarded.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "5d861be8c87b8a8f8e0f58b3b8864f31da7ee9c3", "tree": "79696e03f05424e4ced522f40435e8dfed13dfc7", "parents": [ "1d84612649427a85e1f311baa7215f9a6252d856" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Thu Jun 28 19:16:51 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 21:50:53 2007 -0700" }, "message": "IB/cm: cm_msgs.h should include ib_cm.h\n\ncm_msgs.h uses definitions from ib_cm.h. Include it directly, rather\nthan depending on a specific include order.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "1d84612649427a85e1f311baa7215f9a6252d856", "tree": "cbb9d614cb56c12f76e438f3f05ca9559a2dee3a", "parents": [ "24be6e81c78314c91a47200272eb4bc31284bd7b" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Mon Jun 18 11:09:37 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 21:50:05 2007 -0700" }, "message": "IB/cm: Include HCA ACK delay in local ACK timeout\n\nThe IB CM should include the HCA ACK delay when calculating the local\nACK timeout value to use for RC QPs. If the HCA ACK delay is large\nenough relative to the packet life time, then if it is not taken into\naccount, the calculated timeout value ends up being too small, which\ncan result in \"retry exceeded\" errors.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "24be6e81c78314c91a47200272eb4bc31284bd7b", "tree": "bbdc6d8789ae5e2be3dad7930b3eacbbb1dcddec", "parents": [ "2aec5c602c6a44e2a3a173339a9ab94549658e4b" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Mon Jun 18 11:09:36 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 21:47:29 2007 -0700" }, "message": "IB/cm: Use spin_lock_irq() instead of spin_lock_irqsave() when possible\n\nThe ib_cm is a little over zealous about using spin_lock_irqsave,\nwhen spin_lock_irq would do.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "2aec5c602c6a44e2a3a173339a9ab94549658e4b", "tree": "b16a5002a07d96b88ad162ac34d628b531844a6f", "parents": [ "1b844afe9e67d6cd441ae6df71051b4004f31dd2" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Mon Jun 18 11:03:58 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 21:45:31 2007 -0700" }, "message": "IB/sa: Make sure SA queries use default P_Key\n\nMADs sent to the SA should use the the default P_Key (0x7fff/0xffff).\nThere\u0027s no requirement that the default P_Key is stored at index 0 in\nthe local P_Key table, so add code to the sa_query module to look up\nthe index of the default P_Key when creating an address handle for the\nSA (which is done any time the P_Key table might change), and use this\nindex for all SA queries.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "1b844afe9e67d6cd441ae6df71051b4004f31dd2", "tree": "4242f07678723cded8e36bd3e4e60d926e08647e", "parents": [ "8909c571fa1e62e254c4045394e6eaccfadec6f4" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 13:43:53 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 13:43:53 2007 -0700" }, "message": "IPoIB: Recycle loopback skbs instead of freeing and reallocating\n\nInfiniBand HCAs replicate multicast packets back to the QP that sent\nthem if that QP is attached to the destination multicast group. This\nmeans that IPoIB multicasts are often replicated back to the receive\nqueue of the interface that generated them. To avoid confusing the\nnetwork stack, we drop these duplicates within the IPoIB driver.\n\nHowever, there\u0027s no reason to free the skb that received the duplicate\nand then immediately allocate a new skb to post to the receive queue.\nWe can be more efficient and just repost the same skb.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8909c571fa1e62e254c4045394e6eaccfadec6f4", "tree": "aa1a94dd456f05242cec5b47f49183160994cb15", "parents": [ "20089ca55786a243c7b72becd1bf670f4e2c2028" ], "author": { "name": "Shani Moideen", "email": "shani.moideen@wipro.com", "time": "Mon Jun 18 08:46:41 2007 +0530" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 12:28:05 2007 -0700" }, "message": "IB/mthca: Replace memset(\u003caddr\u003e, 0, PAGE_SIZE) with clear_page(\u003caddr\u003e)\n\nSigned-off-by: Shani Moideen \u003cshani.moideen@wipro.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n----\n\n" }, { "commit": "20089ca55786a243c7b72becd1bf670f4e2c2028", "tree": "76b0c517dfa9b4c74ccafac434e4095c7a836e45", "parents": [ "856c52a741950dc0be2c5c231efec626e9a0a3fa" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 11:18:34 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 11:18:34 2007 -0700" }, "message": "IPoIB/cm: Fix warning if IPV6 is not enabled\n\nFix\n\n drivers/infiniband/ulp/ipoib/ipoib_cm.c:1151: warning: unused variable \u0027dev\u0027\n\nby getting rid of the variable dev, which is only used if CONFIG_IPV6\nis enabled, and replacing the one use of it with the value it is\nassigned, namely priv-\u003edev.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "856c52a741950dc0be2c5c231efec626e9a0a3fa", "tree": "9cb262566a3a0521ce6b560813443b3f19b744a4", "parents": [ "f72d2f081453d8b1ea36e47478c3463042746ddc" ], "author": { "name": "Dotan Barak", "email": "dotanb@dev.mellanox.co.il", "time": "Tue Jul 10 16:55:57 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 10 11:04:40 2007 -0700" }, "message": "IB/core: Take sizeof the correct pointer when calling kmalloc()\n\nWhen allocating out_mad in show_pma_counter(), take sizeof *out_mad\ninstead of sizeof *in_mad. It is true that today the type of in_mad\nand out_mad are the same, but this patch will give us a cleaner code.\n\nSigned-off-by: Dotan Barak \u003cdotanb@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "f72d2f081453d8b1ea36e47478c3463042746ddc", "tree": "add8200b720656a1dc36f165f5e8872e5db30633", "parents": [ "8705ce5b90118be93eb8b0ed6f49ca5ff377df24" ], "author": { "name": "Hoang-Nam Nguyen", "email": "hnguyen@de.ibm.com", "time": "Mon Jul 09 15:33:52 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Improve latency by unlocking after triggering the hardware\n\nKick the hardware before unlocking the send/receive queue to overlap\nprocessing a little more.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8705ce5b90118be93eb8b0ed6f49ca5ff377df24", "tree": "a7d273e77e01a4be09882aae2a8285e31e02aebf", "parents": [ "b1cfe43d4b24144596d277133e0e5b715eea1347" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:32:22 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Notify consumers of LID/PKEY/SM changes after nondisruptive events\n\nWhen firmware reports a nondisruptive port configuration change event,\nprevious versions of the eHCA driver didn\u0027t forward the event to consumers\nlike IPoIB. Add code that determines the type of configuration change by\ncomparing old and new port attributes and reports it.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b1cfe43d4b24144596d277133e0e5b715eea1347", "tree": "1e18ebc7aac92ed6430b94a5bb980c01b263f1fd", "parents": [ "26ed687fdd541c2542b79dcd75fb2c82eb36f189" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:31:53 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Return QP pointer in poll_cq()\n\nAlso add two unlikely() statements.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "26ed687fdd541c2542b79dcd75fb2c82eb36f189", "tree": "45d5ad42b6b9259f273cc274d0fbf18a6d169e91", "parents": [ "28db6beb420c756c61dd44d9f2786a0677159e74" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:31:10 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Change idr spinlocks into rwlocks\n\nThis eliminates lock contention among IRQs as well as the need to\ndisable IRQs around idr_find, because there are no IRQ writers.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "28db6beb420c756c61dd44d9f2786a0677159e74", "tree": "3d5b7cfdfcfe80268bf47d84404d2d9a2034cc65", "parents": [ "9844b71baa60270110eabaa9589d3260443d1a71" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:30:39 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Refactor sync between completions and destroy_cq using atomic_t\n\n- ehca_cq.nr_events is made an atomic_t, eliminating a lot of locking.\n- The CQ is removed from the CQ idr first now to make sure no more\n completions are scheduled on that CQ. The \"wait for all completions to\n end\" code becomes much simpler this way.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "9844b71baa60270110eabaa9589d3260443d1a71", "tree": "30e443f3329bf1c0b4e78d0787c8e33158dff59e", "parents": [ "15f001ec47b049051f679f8b8c965ac9aae95b3e" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:29:03 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Lock renaming, static initializers\n\n- Rename all spinlock flags to \"flags\", matching the vast majority of kernel\n code.\n- Move hcall_lock into the only file it\u0027s used in.\n- Replaced spin_lock_init() and friends with static initializers for\n global variables.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "15f001ec47b049051f679f8b8c965ac9aae95b3e", "tree": "1b597102f72cd6d74e37bd8c0febede50e1e92cd", "parents": [ "85f003172fc130626261730141ed021258e81f85" ], "author": { "name": "Hoang-Nam Nguyen", "email": "hnguyen@de.ibm.com", "time": "Mon Jul 09 15:28:18 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Report RDMA atomic attributes in query_qp()\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "85f003172fc130626261730141ed021258e81f85", "tree": "155518aee49caeaf00e686a4c662bea2061f6caf", "parents": [ "472803dab844c0a8a5d757d4c67fa5e76013dcbd" ], "author": { "name": "Stefan Roscher", "email": "stefan.roscher@de.ibm.com", "time": "Mon Jul 09 15:27:13 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Set SEND_GRH flag for all non-LL UD QPs on eHCA2\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "472803dab844c0a8a5d757d4c67fa5e76013dcbd", "tree": "7c0bec5cb66cfd76ad025ca5d4d7a30f613eb7df", "parents": [ "a6a12947fbf4a1782535468d756b0d44babf9760" ], "author": { "name": "Stefan Roscher", "email": "stefan.roscher@de.ibm.com", "time": "Mon Jul 09 15:26:31 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: Support UD low-latency QPs\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "a6a12947fbf4a1782535468d756b0d44babf9760", "tree": "4819ad9abd03f92e51c076745e7068028d06b105", "parents": [ "9a79fc0a1b815cbd05a8e37ea838acfccb7235cc" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:25:10 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: add Shared Receive Queue support\n\nSupport SRQs on eHCA2. Since an SRQ is a QP for eHCA2, a lot of code\n(structures, create, destroy, post_recv) can be shared between QP and SRQ.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "9a79fc0a1b815cbd05a8e37ea838acfccb7235cc", "tree": "ed94fc291a5017f94e2b9d57f6909a84e1eb40db", "parents": [ "91f13aa3fc22e357b494c5b8270e94543870928d" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:23:15 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: QP code restructuring in preparation for SRQ\n\n- Replace init_qp_queues() by a shorter init_qp_queue(), eliminating\n duplicate code.\n\n- hipz_h_alloc_resource_qp() doesn\u0027t need a pointer to struct ehca_qp any\n longer. All input and output data is transferred through the parms\n parameter.\n\n- Change the interface to also support SRQ.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "91f13aa3fc22e357b494c5b8270e94543870928d", "tree": "e8f78c2549945f5e3bc42e33ce81a18a9292f746", "parents": [ "b8a3ba551369982180917a999d32fcedbba34115" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Mon Jul 09 15:21:45 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:27 2007 -0700" }, "message": "IB/ehca: HW level, HW caps and MTU autodetection\n\nIn preparation for support of new eHCA2 features, change adapter probing:\n - Hardware level is changed to encode major and minor chip version\n - Hardware capabilities are queried from the firmware\n - The maximum MTU is queried from the firmware instead of assuming a\n fixed value\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b8a3ba551369982180917a999d32fcedbba34115", "tree": "5409690f02ab7c9c1b373021ee6c146025f4c914", "parents": [ "f41d229865c984015914221959675b1c8723f6a7" ], "author": { "name": "Hoang-Nam Nguyen", "email": "hnguyen@de.ibm.com", "time": "Mon Jul 09 15:20:55 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ehca: Change scaling_code parameter description to match default value\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "f41d229865c984015914221959675b1c8723f6a7", "tree": "c5268dbcf6ea6ea2fc3af9387559d6ff6a265da1", "parents": [ "78526821bee8506ab903f596e8e6187fa5919877" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Thu Jun 28 19:16:20 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: return correct PortGUID in NodeInfo\n\nReturn the PortGUID of the correct port when responding to a NodeInfo\nquery. Returning the SystemImageGUID causes issues when there are\nmultiple HCAs in a single system.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "4f5973fd3bed9dbff0940ad6cc918be64a17b3e4", "tree": "1d5a5f1433da4138e228d9cbfddf926a7daf5d46", "parents": [ "3588423fbab1cfaf839e67378897e232a054f317" ], "author": { "name": "Arthur Jones", "email": "arthur.jones@qlogic.com", "time": "Fri Jul 06 12:48:58 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Remove bogus RD_ATOMIC checks from modify_qp\n\nThe changeset 3859e39d (\"IB/ipath: Support larger IB_QP_MAX_DEST_RD_ATOMIC\nand IB_QP_MAX_QP_RD_ATOMIC\") added support for larger RD_ATOMIC values,\nbut it failed to take out the stricter checks that were before these and\nhence had no effect. This patch takes out the bogus checks....\n\nSigned-off-by: Arthur Jones \u003carthur.jones@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "3588423fbab1cfaf839e67378897e232a054f317", "tree": "4885593a70ef1168a33ec261b1e0fad7f7a6a11c", "parents": [ "9ca48655667214be6ebd191628a3c4b5b529a87e" ], "author": { "name": "Arthur Jones", "email": "arthur.jones@qlogic.com", "time": "Fri Jul 06 12:48:53 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Test interrupts at driver startup\n\nAll too often, interrupts do not get enabled for our card due to BIOS\nmisconfiguration and other issues. This patch checks for that\ncondition on startup and warns the user. This patch is based on work\n(check LID availability) by Robert Walsh.\n\nSigned-off-by: Arthur Jones \u003carthur.jones@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "9ca48655667214be6ebd191628a3c4b5b529a87e", "tree": "89670425b57296ebeeb0ca86d818e36efd8c0603", "parents": [ "12f9a49e1bce241c243268c764e1b37391157cd0" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Fri Jul 06 12:48:48 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Remove support for preproduction HTX InfiniPath cards\n\nClean up some code by removing support for some older pre-production\nHTX InfiniPath cards.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\n" }, { "commit": "12f9a49e1bce241c243268c764e1b37391157cd0", "tree": "e93eeb98aa51a4f41d7888a69d4216bb255abdd5", "parents": [ "37a7e9b7f28fbef4b6abda102fa41c1467f6022f" ], "author": { "name": "Dave Olson", "email": "dave.olson@qlogic.com", "time": "Fri Jul 06 12:48:43 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Change version wording to be less confusing with release number\n\nSigned-off-by: Dave Olson \u003cdave.olson@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "37a7e9b7f28fbef4b6abda102fa41c1467f6022f", "tree": "a11e49052ffbf12c124477c1239e2bd4363054a2", "parents": [ "f42b6471e9e9082ed4bcba1456eccadf98f1337a" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Fri Jul 06 12:48:38 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Lower default number of kernel send buffers\n\nThe default calculation for the number of send buffers to allocate to\nthe kernel was too high for the PCIe version of the chip thus leaving\nfewer than desired send buffers for user MPI applications.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "0f4fc5ebd9a68ede20ca365576a6df2df2fefc4c", "tree": "0b41785bfd9ca49fa365d1b34ae2c81bce5c3caf", "parents": [ "4fc570bcbe77f823aae183dd824869f79e74cc97" ], "author": { "name": "Dave Olson", "email": "dave.olson@qlogic.com", "time": "Fri Jul 06 12:48:33 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Be more cautious about coming out of freeze mode\n\nWe are more careful to be sure that we don\u0027t lose information about\nchanges that occurred while we were in freeze mode, when the chip will\nnot notify us, and try to avoid false error interrupts while doing\ncleanup. Put all of this logic in a new function ipath_clear_freeze().\n\nSigned-off-by: Dave Olson \u003cdave.olson@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "4fc570bcbe77f823aae183dd824869f79e74cc97", "tree": "3955946608a973b06db8b7e5b3ec319354a6fa73", "parents": [ "06cc85086e6896939f8c68f8518224748f6b0b2f" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Fri Jul 06 12:48:23 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Add barrier before updating WC head in shared memory\n\nAdd a barrier to make sure the CPU doesn\u0027t reorder writes to memory,\nsince user programs can be polling on the head index update and the\nentry should be written before that.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "06cc85086e6896939f8c68f8518224748f6b0b2f", "tree": "7c568a68748814a7baa85d0d8d86edd66f16fd97", "parents": [ "6abb6ea80b6564c00906aa0727cf59d2be780d82" ], "author": { "name": "Jan Engelhardt", "email": "jengelh@linux01.gwdg.de", "time": "Wed May 23 14:21:22 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB: Use menuconfig for InfiniBand menu\n\nChange Kconfig objects from \"menu, config\" into \"menuconfig\" so\nthat the user can disable the whole feature without having to\nenter the menu first.\n\nSigned-off-by: Jan Engelhardt \u003cjengelh@gmx.de\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "6abb6ea80b6564c00906aa0727cf59d2be780d82", "tree": "dbfd28b4c8318706bea801cb72da294ed826e221", "parents": [ "ecc2f0060fa7ff2fc53864ee19e370e5ddd47d5e" ], "author": { "name": "WANG Cong", "email": "xiyou.wangcong@gmail.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: Check return of kmalloc() in iwch_register_device()\n\nSigned-off-by: WANG Cong \u003cxiyou.wangcong@gmail.com\u003e\n[ Also remove cast from void * return of kmalloc() as suggested by \n Jesper Juhl \u003cjesper.juhl@gmail.com\u003e. ]\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "ecc2f0060fa7ff2fc53864ee19e370e5ddd47d5e", "tree": "1c6e8e79aeb72b44020e4d1885061a2bd5a62c91", "parents": [ "96d0e4931e264012f57a2ae8f7c4697bfa55386a" ], "author": { "name": "Steve Wise", "email": "swise@opengridcomputing.com", "time": "Mon Jun 25 12:46:00 2007 -0500" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: Don\u0027t abort after failures sending the mpa reply\n\nThis bug results in an abort request being sent down _after_ the tid\nhas been released. If the tid happens to have been reused, then the\nsubsequent generation of the tid gets incorrectly aborted.\n\nThe thread running iwch_accecpt_cr() must not abort a connection if an\nerror is returned after being awakened. If any errors did occur while\niwch_accept_cr() is blocked, then the connection has already been\naborted on the thread processing the error.\n\nSigned-off-by: Steve Wise \u003cswise@opengridcomputing.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "96d0e4931e264012f57a2ae8f7c4697bfa55386a", "tree": "f5b5d0fb3f7bf3897ea0ac16575929b85ef4cdac", "parents": [ "6eda48d1e82f901d977a405e43a849dc79968d61" ], "author": { "name": "Steve Wise", "email": "swise@opengridcomputing.com", "time": "Thu Jun 21 18:17:57 2007 -0500" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: Don\u0027t post TID_RELEASE message\n\nThe LLD does this for us in cxgb3_remove_tid().\n\nAlso fixed active open failure cases where we also shouldn\u0027t be\nreleasing the TID.\n\nSigned-off-by: Steve Wise \u003cswise@opengridcomputing.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "6eda48d1e82f901d977a405e43a849dc79968d61", "tree": "d1131347cde7663d44a98c22ace38616a657a5fe", "parents": [ "1580367e7b2068d075cd42d04c4b8c274815e6fc" ], "author": { "name": "Steve Wise", "email": "swise@opengridcomputing.com", "time": "Tue Jun 19 09:27:48 2007 -0500" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: ctrl-qp init/clear shouldn\u0027t set the gen bit\n\nSigned-off-by: Steve Wise \u003cswise@opengridcomputing.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "1580367e7b2068d075cd42d04c4b8c274815e6fc", "tree": "594ac80b4d09207776db6ea7ab81b92fc83e469e", "parents": [ "fb497d726612bb0e50240405aaabcf37f13901ed" ], "author": { "name": "Steve Wise", "email": "swise@opengridcomputing.com", "time": "Tue Jun 19 09:27:48 2007 -0500" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: Don\u0027t count neg_adv abort_req_rss messages as real aborts\n\nNegative advice messages should _not_ count toward the 2 abort\nrequests needed to indicate an abort request.\n\nSigned-off-by: Steve Wise \u003cswise@opengridcomputing.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "fb497d726612bb0e50240405aaabcf37f13901ed", "tree": "2c6e6cec4e1de3e8012cd02075c401d927209a04", "parents": [ "de3d353072f9342f04112ba0504c3e294220cb8f" ], "author": { "name": "Steve Wise", "email": "swise@opengridcomputing.com", "time": "Tue Jun 19 09:27:48 2007 -0500" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: TERMINATE WRs can hang the tx ofld queue\n\nDon\u0027t set the gen bits nor length bits in the terminate WR. This is\ndone by the LLD driver.\n\nSigned-off-by: Steve Wise \u003cswise@opengridcomputing.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "de3d353072f9342f04112ba0504c3e294220cb8f", "tree": "94c07283ea583f72544eb76795aa127bef6c5aa9", "parents": [ "149983af609e8f5c57157467baf8545d17b8a6a1" ], "author": { "name": "Steve Wise", "email": "swise@opengridcomputing.com", "time": "Mon May 14 13:27:27 2007 -0500" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "RDMA/cxgb3: Streaming -\u003e RDMA mode transition fixes\n\nDue to a HW issue, our current scheme to transition the connection from\nstreaming to rdma mode is broken on the passive side. The firmware\nand driver now support a new transition scheme for the passive side:\n\n - driver posts rdma_init_wr (now including the initial receive seqno)\n - driver posts last streaming message via TX_DATA message (MPA start\n response)\n - uP atomically sends the last streaming message and transitions the\n tcb to rdma mode.\n - driver waits for wr_ack indicating the last streaming message was ACKed.\n\nNOTE: This change also bumps the required firmware version to 4.3.\n\nSigned-off-by: Steve Wise \u003cswise@opengridcomputing.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "149983af609e8f5c57157467baf8545d17b8a6a1", "tree": "4a8ea99e670a6e74b7b03122408a82d4c6e29518", "parents": [ "87427da55bc03dbce7906a5b09ed50279d654d28" ], "author": { "name": "Dotan Barak", "email": "dotanb@dev.mellanox.co.il", "time": "Tue Jun 26 15:55:28 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "mlx4_core: Get the maximum message size from reported device capabilities\n\nGet the maximum message size from the device capabilities returned\nfrom the QUERY_DEV_CAP firmware command, rather than hard-coding 2 GB.\n\nSigned-off-by: Dotan Barak \u003cdotanb@dev.mellanox.co.il\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n\n" }, { "commit": "87427da55bc03dbce7906a5b09ed50279d654d28", "tree": "e1ad186525a0c54a6e90fcc88bd87d0da5093e13", "parents": [ "991bda284dcbc1ed0522683320043ac41d70c82c" ], "author": { "name": "John Gregor", "email": "john.gregor@qlogic.com", "time": "Mon Jun 11 10:21:14 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Update copyright dates\n\nNow that it\u0027s June, it\u0027s about time to update\nthe copyright notices of files that have changed.\n\nSigned-off-by: John Gregor \u003cjohn.gregor@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "991bda284dcbc1ed0522683320043ac41d70c82c", "tree": "d35f8804a7ccc06ad30d2b87fcce56d5b0ca79f1", "parents": [ "f2d042313e420002b06715675963cfab48ed2597" ], "author": { "name": "Robert Walsh", "email": "robert.walsh@qlogic.com", "time": "Mon Jun 04 09:55:48 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Clean send flags properly on QP reset\n\nSigned-off-by: Robert Walsh \u003crobert.walsh@qlogic.com\u003e\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "f2d042313e420002b06715675963cfab48ed2597", "tree": "ae149c8d21220aff9157952e39a7632be8c7ba61", "parents": [ "b506e1dc59726a1c608f26e7294b9fe186255139" ], "author": { "name": "Robert Walsh", "email": "robert.walsh@qlogic.com", "time": "Mon Jun 18 14:24:49 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: ipath_poll fixups and enhancements\n\nFix ipath_poll and enhance it so we can poll for urgent packets or\nregular packets and receive notifications of when a header queue\noverflows.\n\nSigned-off-by: Robert Walsh \u003crobert.walsh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b506e1dc59726a1c608f26e7294b9fe186255139", "tree": "ae8550e392380cb953e19ab92ba0ae0716bcc1ab", "parents": [ "e8e7ad711509f576b1bffd92c3ae4672fe92ec48" ], "author": { "name": "Robert Walsh", "email": "robert.walsh@qlogic.com", "time": "Mon Jun 18 14:24:48 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Send ACK invalid where appropriate\n\nThe IB specification ch. 9.9.3 table 58 says that a QP which isn\u0027t set\nup for the operation should return a NAK invalid request.\n\nSigned-off-by: Robert Walsh \u003crobert.walsh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e8e7ad711509f576b1bffd92c3ae4672fe92ec48", "tree": "bdb211849ca0354632882161ff46647bc4a1f015", "parents": [ "bacf4013530e7fc230a8aa0c6ea3c17fc2f47665" ], "author": { "name": "Michael Albaugh", "email": "Michael.Albaugh@qlogic.com", "time": "Mon Jun 18 14:24:47 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Add capability to modify PBC word\n\nDuring compliance testing and when debugging some interconnect issues,\nit is very useful to be able to send malformed packets, without having\nthe device signal them as malformed (drop, or terminate with EBP). The\nhardware supports this, but the driver \"diagnostic packet\" interface\ndid not.\n\nExtend capability to send specific malformed packets for testing.\n\nSigned-off-by: Michael Albaugh \u003cMichael.Albaugh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "bacf4013530e7fc230a8aa0c6ea3c17fc2f47665", "tree": "29ac5cc3512b87df9be1e074348fe0ee2809d17c", "parents": [ "0df6291c8af2778d05f278d5738eef2c8fafa2dd" ], "author": { "name": "Mark Debbage", "email": "mark.debbage@qlogic.com", "time": "Mon Jun 18 14:24:46 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Make handling of one subport consistent\n\nPreviously the driver and userspace code handled the case of 1 subport\nsomewhat inconsistently. The new interpretation of this situation is\nthat if one subport is requested, the driver turns on the subport\nmechanism and arranges for the port to be \"shared\" by one process. In\nnormal use the userspace library does not use this configuration and\ninstead arranges for the port not to be shared at all. This\nparticular idiom can be useful for testing purposes.\n\nSigned-off-by: Mark Debbage \u003cmark.debbage@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "0df6291c8af2778d05f278d5738eef2c8fafa2dd", "tree": "07eb25e1da424197dc89b0ace53d7c366b271cfa", "parents": [ "d781b129f1e8b3e2f369d8035a61a5233832e65c" ], "author": { "name": "Mark Debbage", "email": "mark.debbage@qlogic.com", "time": "Mon Jun 18 14:24:45 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Correct checking of swminor version field when using subports\n\nWhen subports are required to run a program, this patch checks that\nthe driver and the userspace library have compatible subport\nimplementations. This is achieved through checks on the swminor\nversion field built into the driver and userspace library. Bad\ncombinations are reported through syslog and result in an error when\nopening the port.\n\nSigned-off-by: Mark Debbage \u003cmark.debbage@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "d781b129f1e8b3e2f369d8035a61a5233832e65c", "tree": "7ff0bfbcd2e32a4693fa2007b773a7ca87968464", "parents": [ "30d149ab58cc3ed8e4bc9c4dc45bebbed0e84b6e" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:44 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Duplicate RDMA reads can cause responder to NAK inappropriately\n\nA duplicate RDMA read request can fool the responder into NAKing a new\nRDMA read request because the responder wasn\u0027t keeping track of\nwhether the queue of RDMA read requests had been sent at least once.\nFor example, requester sends 4 2K byte RDMA read requests, times out,\nand resends the first, then sees the 4 responses, then sends a 5th\nRDMA read or atomic operation. The responder sees the 4 requests,\nsends 4 responses, sees the resent 1st request, rewinds the queue,\nthen sees the 5th request but thinks the queue is full and that the\nrequester is invalidly sending a 5th new request.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "30d149ab58cc3ed8e4bc9c4dc45bebbed0e84b6e", "tree": "f85f47458efb9e98f01b490a539dbf873bbaddd9", "parents": [ "db5518cd09c21f0fa70af0a4ca38badd90622c9e" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:44 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Fix possible data corruption if multiple SGEs used for receive\n\nThe code to copy data from the receive queue buffers to the IB SGEs\ndoesn\u0027t check the SGE length, only the memory region/page length when\ncopying data. This could overwrite parts of the user\u0027s memory that\nwere not intended to be written. It can only happen if multiple SGEs\nare used to describe a receive buffer which almost never happens in\npractice.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "db5518cd09c21f0fa70af0a4ca38badd90622c9e", "tree": "c8183dc863a03411bdb2f8737870ea3ed743d8c8", "parents": [ "06ee109002672ac875558ec699b53cf08a865bd3" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:43 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Wait for PIO available interrupt\n\nThe send function is called when posting new send work requests.\nThere is no point in trying to send a packet if the QP is already\nwaiting for a HW send buffer so don\u0027t clear the busy bit until the\nbuffer available interrupt happens.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "06ee109002672ac875558ec699b53cf08a865bd3", "tree": "6d76bae2f40731b27d6b11f71c0cd9ff200e45af", "parents": [ "9380068fc2f230e7840ff87d3f1e6030ae2ee5e8" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:42 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Fix RDMA read retry code\n\nA RDMA read response or atomic response can ACK earlier sends and RDMA\nwrites. In this case, the wrong work request pointer was being used\nto store the read first response or atomic result. Also, if a RDMA\nread request is retried, the code to compute which request to resend\nwas incorrect.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "9380068fc2f230e7840ff87d3f1e6030ae2ee5e8", "tree": "693349ae9f8dc25156091f7834030b35e01f23e1", "parents": [ "561095f20eeb5c6c05c303bad8cdb3f8a80821bc" ], "author": { "name": "Dave Olson", "email": "dave.olson@qlogic.com", "time": "Mon Jun 18 14:24:41 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Use S_ABORT not cancel and abort on exit freeze mode after recovery\n\nThis centralizes the use of the abort functionality, removes the\nunneeded buffer cancel (abort does the same thing), sets up to ignore\nlaunch errors after abort, same as cancel. We need abort on exit from\nfreeze mode to avoid having buffers stuck in the busy state, if a user\nprocess happened to complete the send while we were in freeze mode\ndoing the recovery.\n\nSigned-off-by: Dave Olson \u003cdave.olson@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "561095f20eeb5c6c05c303bad8cdb3f8a80821bc", "tree": "57a8c7e69151ae5c94d73cb1ecfcef1441270792", "parents": [ "f716cdfe57f217966f41a7add190d6f5b9fd0769" ], "author": { "name": "Dave Olson", "email": "dave.olson@qlogic.com", "time": "Mon Jun 18 14:24:40 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Fix the mtrr_add args for chips with 2 buffer sizes\n\nThe values passed have never been right for iba 6120 chips, but just\nhappened to work. We needed to select the right buffer offset in the\nchip (both are in same register), and the total length was wrong also,\nbut was covered by the rounding up.\n\nSigned-off-by: Dave Olson \u003cdave.olson@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "f716cdfe57f217966f41a7add190d6f5b9fd0769", "tree": "fbcbdbddbbcea403d52ee641d2a63b74ce1e9f2f", "parents": [ "1dd6a1be1416be48cafda9e63a614f26f0428d10" ], "author": { "name": "Joan Eslinger", "email": "joan.eslinger@qlogic.com", "time": "Mon Jun 18 14:24:39 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Change use of constants for TID type to defined values\n\nDefine pkt rcvd \u0027type\u0027 in a way consistent with HW spec and chips.\n\nThe hardware considers received packets of type 0 to be expected, and\ntype 1 to be eager. The driver was calling the ipath_f_put_tid\nfunctions using a variable called \u0027type\u0027 set to 0 for eager and to 1\nfor expected packets. Worse, the iba6110 and iba6120 drivers used\nthose values inconsistently. This was quite confusing. Now\neverything is consistent with the hardware.\n\nSigned-off-by: Dave Olson \u003cdave.olson@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "1dd6a1be1416be48cafda9e63a614f26f0428d10", "tree": "a20e33515a80c2da5153d66754cb194503b9974d", "parents": [ "6d2fad0472ca0d6caba7c36d2823a527e2a0e4f5" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:38 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Set M bit in BTH according to IB spec\n\nAccording to chapter 17.2.8.1.1, QPs start in the migrated state and\nshould send packets with the M bit set in the BTH.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "6d2fad0472ca0d6caba7c36d2823a527e2a0e4f5", "tree": "ba9c64425e176f533bd8f33821916b205a2553e6", "parents": [ "2c9749c3b534ea0e606b7ee2c29849bbb8d5b0a9" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:38 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Fix local loopback bug when waiting for resources\n\nThis patch fixes a minor bug where the wrong QP was checked for a send\nwork request that should wait for an RNR timeout.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "2c9749c3b534ea0e606b7ee2c29849bbb8d5b0a9", "tree": "50e4bb91927838c2184d2e90240168096bf234f1", "parents": [ "fdc7215fbd7b7652b052d1fb7893afe324dba7aa" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Mon Jun 18 14:24:37 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Fix problem with next WQE after a UC completion\n\nThis patch fixes a bug introduced when moving some code around for\nreadability.\n\nSetting the wqe pointer at the end of the function is a NOP since it\nisn\u0027t used. Move it back to where it is used.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "fdc7215fbd7b7652b052d1fb7893afe324dba7aa", "tree": "9aff6d20fc7e5abfa2121913357c52b1bf7c8e6c", "parents": [ "e7340f04426416a6655ffaead4651bfb9e1b0848" ], "author": { "name": "Robert Walsh", "email": "robert.walsh@qlogic.com", "time": "Mon Jun 18 14:24:36 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:26 2007 -0700" }, "message": "IB/ipath: Fill in some missing FMR-related fields in query_device\n\nIn ipath_query_device(), some of the struct ib_device_attr fields were\nnot being initialized.\n\nSigned-off-by: Robert Walsh \u003crobert.walsh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e7340f04426416a6655ffaead4651bfb9e1b0848", "tree": "1ce7d3c405bc9e4ea69871e40504b057eeb30e66", "parents": [ "380bf5d38f3cc2799ed2fae554f7af1c4b0ed35b" ], "author": { "name": "Robert Walsh", "email": "robert.walsh@qlogic.com", "time": "Mon Jun 18 14:24:35 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Fix maximum MTU reporting\n\nAlthough our chip supports 4K MTUs, our driver doesn\u0027t yet support\nthis feature, so limit the maximum MTU to 2K until we get support for\n4K MTUs implemented.\n\nSigned-off-by: Robert Walsh \u003crobert.walsh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "380bf5d38f3cc2799ed2fae554f7af1c4b0ed35b", "tree": "fa6627d2a0f93b86cd44c967f5408dd78126729f", "parents": [ "aecd3b5ab19624ca9644b9df9c61615282d8923f" ], "author": { "name": "Dave Olson", "email": "dave.olson@qlogic.com", "time": "Mon Jun 18 14:24:35 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Support the IBA6110 revision 4\n\nRecognize IBA 6110 Revision 4: same feature set, etc. as earlier revisions.\n\nSigned-off-by: Dave Olson \u003cdave.olson@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "aecd3b5ab19624ca9644b9df9c61615282d8923f", "tree": "d4f4b5367a40bd737899b7e3dbfcfab6021891e0", "parents": [ "8e9ab3f1c9e34d5c28446c3738983d33a3937fe0" ], "author": { "name": "Michael Albaugh", "email": "michael.albaugh@qlogic.com", "time": "Thu May 17 07:26:28 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Log \"active\" time and some errors to EEPROM\n\nWe currently track various errors, now we enhance that capability by\nlogging some of them to EEPROM. We also now log a cumulative \"active\"\ntime defined by traffic though the InfiniPath HCA beyond the normal SM\ntraffic.\n\nSigned-off-by: Michael Albaugh \u003cmichael.albaugh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8e9ab3f1c9e34d5c28446c3738983d33a3937fe0", "tree": "edf8e4c3af761c440068b1d1039ab402650f372f", "parents": [ "17b2eb9fe6bfadcb3ece308ed50193d10b71ba6e" ], "author": { "name": "John Gregor", "email": "john.gregor@qlogic.com", "time": "Thu May 17 08:15:50 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Remove incompletely implemented ipath_runtime flags and code\n\nThe IPATH_RUNTIME_PBC_REWRITE and the IPATH_RUNTIME_LOOSE_DMA_ALIGN\nflags were not ever implemented correctly and did not turn out to be\nnecessary. Remove the last vestiges of these flags but mark the spot\nwith a comment to remind us to not reuse these flags in the interest\nof binary compatibility. The INFINIPATH_XGXS_SUPPRESS_ARMLAUNCH_ERR\nbit was also not found to be useful, so it was dropped in the cleanup\nas well.\n\nSigned-off-by: John Gregor \u003cjohn.gregor@qlogic.com\u003e\nSigned-off-by: Arthur Jones \u003carthur.jones@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "17b2eb9fe6bfadcb3ece308ed50193d10b71ba6e", "tree": "567c1e7d11cc14990ecabd11fe23b1941a2ebcfd", "parents": [ "82466f00ec6ef0a5ca7ea8991c731af2ec561c7d" ], "author": { "name": "Michael Albaugh", "email": "michael.albaugh@qlogic.com", "time": "Thu May 17 07:05:04 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Lock and always use shadow copies of GPIO register\n\nThe new LED blinking interface adds more contention for the\nunprotected GPIO pins that were already shared, though not commonly at\nthe same time. We add locks to the accesses to these pins so that\nRead-Modify-Write is now safe. Some of these locks are added at\ninterrupt context, so we shadow the registers which drive and inspect\nthese pins to avoid the mmio read/writes. This mitigates the effects\nof the locks and hastens us through the interrupt.\n\nAdd locking and always use shadows for registers controlling GPIO pins\n(ExtCtrl and GPIOout). The use of shadows implies doing less I/O,\nwhich can make I2C operation too fast on some platforms. An explicit\nudelay(1) in SCL manipulation fixes that.\n\nSigned-off-by: Michael Albaugh \u003cmichael.albaugh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "82466f00ec6ef0a5ca7ea8991c731af2ec561c7d", "tree": "48fcc8f78e6df67130e8c70b160fea42aeff79ca", "parents": [ "a024291b367f98188f4da4a66a9f2f40a2163efb" ], "author": { "name": "Michael Albaugh", "email": "michael.albaugh@qlogic.com", "time": "Wed May 16 15:45:09 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Support blinking LEDs with an led_override file\n\nWhen we want to find an InfiniPath HCA in a rack of nodes, it is often\nexpeditious to blink the status LEDs via a userspace /sys file.\n\nA write-only led_override \"file\" is published per device. Writes to\nthis file are interpreted as (string form) numbers, and the resulting\nvalue sent to ipath_set_led_override(). The upper eight bits are\ninterpretted as a 4.4 fixed-point \"frequency in Hertz\", and the bottom\ntwo 4-bit values are alternately (D0..3, then D4..7) used by the\nboard-specific LED-setting function to override the normal state.\n\nSigned-off-by: Michael Albaugh \u003cmichael.albaugh@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "a024291b367f98188f4da4a66a9f2f40a2163efb", "tree": "a7a0a58a5dad6ce6f6114f8ca5fed9cb9c6fe9ee", "parents": [ "525f5f44c487201c91a11904516ee56dea93644e" ], "author": { "name": "Bryan O\u0027Sullivan", "email": "bryan.osullivan@qlogic.com", "time": "Wed May 16 15:31:19 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:25 2007 -0700" }, "message": "IB/ipath: Include \u003clinux/vmalloc.h\u003e to fix ppc64 build\n\nSigned-off-by: Bryan O\u0027Sullivan \u003cbryan.osullivan@qlogic.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "63019d9329628a5056b3a0c76aee955873dc25c3", "tree": "2dc615e909624961b5883d314ce86260dfe369d7", "parents": [ "1d3f4b905a786d69103d9e6d8e92683fb2c7a027" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Mon Jun 11 08:56:10 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 16:17:33 2007 -0700" }, "message": "IB/mlx4: Include linux/mutex.h from mlx4_ib.h\n\nmlx4_ib.h uses struct mutex, so although \u003clinux/mutex.h\u003e seems to be\npulled in indirectly by one of the headers it includes, the right\nthing is to include \u003clinux/mutex.h\u003e directly.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "1d3f4b905a786d69103d9e6d8e92683fb2c7a027", "tree": "ddb38919eadaf8aff6b59240e0a49f3e85baadc4", "parents": [ "43506d954e43933cd6fdcab679f6ab057e7607c6" ], "author": { "name": "Andrew Morton", "email": "akpm@linux-foundation.org", "time": "Fri Jun 08 16:29:43 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 16:17:33 2007 -0700" }, "message": "IB: Fix ib_umem_get() when npages \u003d\u003d 0\n\ngcc correctly warned:\n\ndrivers/infiniband/core/umem.c: In function \u0027ib_umem_get\u0027:\ndrivers/infiniband/core/umem.c:78: warning: \u0027ret\u0027 may be used uninitialized in this function\n\nSet ret to 0 in case npages \u003d\u003d 0 and the loop isn\u0027t entered at all.\n\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "43506d954e43933cd6fdcab679f6ab057e7607c6", "tree": "8c0f10a44f464cd3193d5c05c1cedbc56e662a0b", "parents": [ "fffba373ef8cc9a5881353395c4edd1ab27cf154" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 16:17:32 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 16:17:32 2007 -0700" }, "message": "IB: Remove garbage non-ASCII characters from comments\n\nA few files had 0xa0 characters in comments. Remove them so that the \nfiles are clean ASCII text.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "fffba373ef8cc9a5881353395c4edd1ab27cf154", "tree": "53b83634f9e9a4978c56c3c03d19fa58182a2556", "parents": [ "1bae4dbf9576e563da23927e4078fffbbce67a75" ], "author": { "name": "Joachim Fenkes", "email": "fenkes@de.ibm.com", "time": "Thu May 24 16:51:05 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 16:17:32 2007 -0700" }, "message": "IB/ehca: Refactor \"maybe missed event\" code\n\nRefactor the ehca changes from commit ed23a727 (\"IB: Return \"maybe\nmissed event\" hint from ib_req_notify_cq()\") so the queue arithmetic\nis done in slightly fewer lines. Also, move the spinlock flags into\nthe block they\u0027re used in.\n\nSigned-off-by: Joachim Fenkes \u003cfenkes@de.ibm.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "1bae4dbf9576e563da23927e4078fffbbce67a75", "tree": "3728d91be2f42c1a4a73e41c92857769738d1b83", "parents": [ "71780f59e127bb281a9302d430495ca9586c14e7" ], "author": { "name": "Hal Rosenstock", "email": "halr@voltaire.com", "time": "Mon May 14 17:21:52 2007 -0400" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 16:17:32 2007 -0700" }, "message": "IB/mad: Enhance SMI for switch support\n\nExtend the SMI with switch (intermediate hop) support. Care has been\ntaken to ensure that the CA (and router) code paths are changed as\nlittle as possible.\n\nSigned-off-by: Suresh Shelvapille \u003csuri@baymicrosystems.com\u003e\nSigned-off-by: Hal Rosenstock \u003chalr@voltaire.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "841adfca9c5fc0fec6b1f0b2e5eb7a3b239a7730", "tree": "fcdc834ad623a890a0eeb499c41593615746c706", "parents": [ "786f238e4f2f0d962831ee3228f2327133eb082a" ], "author": { "name": "Ralph Campbell", "email": "ralph.campbell@qlogic.com", "time": "Fri Jun 29 11:37:56 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 02 20:48:31 2007 -0700" }, "message": "IPoIB/cm: Partial error clean up unmaps wrong address\n\nIf a page can\u0027t be allocated for the frag list of a skb, the code to\nunmap the partially allocated list is off by one. For exaple, if\n\u0027frags\u0027 equals one, i \u003d\u003d 0, and the alloc_page() fails, then the old\nloop would have unmapped mapping[1] which is uninitialized. The same\nwould happen if the call to ib_dma_map_page() failed.\n\nSigned-off-by: Ralph Campbell \u003cralph.campbell@qlogic.com\u003e\nAcked-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "c8681f14013d3ad2fc4fb4e30cfd4ea548f7a249", "tree": "4ea047565dcc3340dcd473fca5579a5cb2de349a", "parents": [ "13ef5f44c3931dff1d75443a875e97b588d4b8f0" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Jun 21 13:39:10 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 13:39:10 2007 -0700" }, "message": "IB/mlx4: Correct max_srq_wr returned from mlx4_ib_query_device()\n\nWe need to keep a spare entry in the SRQ so that there always is a\nnext WQE available when posting receives (so that we can tell the\ndifference between a full queue and an empty queue). So subtract 1\nfrom the value HW gives us before reporting the limit on SRQ entries\nto consumers.\n\nFound by Mellanox QA.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "13ef5f44c3931dff1d75443a875e97b588d4b8f0", "tree": "7db19c7302bf5a0bc851ea55106918456d84b494", "parents": [ "82c3aca6ad9004169df8f2f8c0747686fe4003b3" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 13:39:08 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 13:39:08 2007 -0700" }, "message": "IPoIB/cm: Remove dead definition of struct ipoib_cm_id\n\nIt\u0027s completely unused.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "82c3aca6ad9004169df8f2f8c0747686fe4003b3", "tree": "ab24b09038b432e93e6afdcd660447ca1714eaa6", "parents": [ "3ec7393a6858a1716e74aa81be6af76fd180021d" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Wed Jun 20 19:22:15 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 13:38:08 2007 -0700" }, "message": "IPoIB/cm: Fix interoperability when MTU doesn\u0027t match\n\nIPoIB connected mode currently rejects a connection request unless the\nsupported MTU is \u003e\u003d the local netdevice MTU. This breaks\ninteroperability with implementations that might have tweaked\nIPOIB_CM_MTU, and there\u0027s real no longer a reason to do so: this test\nis just a leftover from when we did not tweak MTU per-connection. Fix\nthis by making the test as permissive as possible.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "3ec7393a6858a1716e74aa81be6af76fd180021d", "tree": "2f797894f98a1321605dbcf2610c6349bd92003a", "parents": [ "24bce5080306bd5255cbda3d6b09a29d5515b470" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Tue Jun 19 13:40:41 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 13:03:50 2007 -0700" }, "message": "IPoIB/cm: Initialize RX before moving QP to RTR\n\nFix a crasher bug in IPoIB CM: once a QP is in the RTR state, a\nreceive completion (or even an asynchronous error) might be observed\non this QP, so we have to initialize all of our receive data\nstructures before moving to the RTR state.\n\nAs an optimization (since modify_qp might take a long time), the\njiffies update done when moving RX to the passive_ids list is also\nleft in place to reduce the chance of the RX being misdetected as\nstale.\n\nThis fixes bug \u003chttps://bugs.openfabrics.org/show_bug.cgi?id\u003d662\u003e.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "24bce5080306bd5255cbda3d6b09a29d5515b470", "tree": "d0061299531d0cda6bd83196f8f4c32c5f93af15", "parents": [ "d025d7858f7415f558e89d870ad1a205954b64cd" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 11:05:58 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 21 11:05:58 2007 -0700" }, "message": "IB/umem: Fix possible hang on process exit\n\nIf ib_umem_release() is called after ib_uverbs_close() sets context-\u003eclosing,\nthen a process can get stuck in a D state, because the code boils down to\n\n\tif (down_write_trylock(\u0026mm-\u003emmap_sem))\n\t\tdown_write(\u0026mm-\u003emmap_sem);\n\nwhich is obviously a stupid instant deadlock. Fix the code so that we\nonly try to take the lock once.\n\nThis bug was introduced in commit f7c6a7b5 (\"IB/uverbs: Export\nib_umem_get()/ib_umem_release() to modules\") which fortunately never\nmade it into a release, and was reported by Pete Wyckoff \u003cpw@osc.edu\u003e.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e61ef2416b0b92828512b6cfcd0104a02b6431fe", "tree": "51d3307aa5be5591f5859f96a3bd1dd20231b9b0", "parents": [ "5ae2a7a836be660ff1621cce1c46930f19200589" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 09:23:47 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 09:23:47 2007 -0700" }, "message": "IB/mlx4: Make sure inline data segments don\u0027t cross a 64 byte boundary\n\nInline data segments in send WQEs are not allowed to cross a 64 byte\nboundary. We use inline data segments to hold the UD headers for MLX\nQPs (QP0 and QP1). A send with GRH on QP1 will have a UD header that\nis too big to fit in a single inline data segment without crossing a\n64 byte boundary, so split the header into two inline data segments.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "5ae2a7a836be660ff1621cce1c46930f19200589", "tree": "655b94b9a016cec92f319761afe6bb3000f5f4fa", "parents": [ "082dee3216c99a838af40be403799f60bcea2e97" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:15:02 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:15:02 2007 -0700" }, "message": "IB/mlx4: Handle FW command interface rev 3\n\nUpcoming firmware introduces command interface revision 3, which\nchanges the way port capabilities are queried and set. Update the\ndriver to handle both the new and old command interfaces by adding a\nnew MLX4_FLAG_OLD_PORT_CMDS that it is set after querying the firmware\ninterface revision and then using the correct interface based on the\nsetting of the flag.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "082dee3216c99a838af40be403799f60bcea2e97", "tree": "4741bc693e9eb882e5c72ab2d206cbe63e9dafaa", "parents": [ "54e95f8dcbd7d86f79b423e8d11053ec9a2d9946" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Mon Jun 18 08:13:59 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:13:59 2007 -0700" }, "message": "IB/mlx4: Handle buffer wraparound in __mlx4_ib_cq_clean()\n\nWhen compacting CQ entries, we need to set the correct value of the\nownership bit in case the value is different between the index we copy\nthe CQE from and the index we copy it to.\n\nFound by Ronni Zimmerman of Mellanox.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "54e95f8dcbd7d86f79b423e8d11053ec9a2d9946", "tree": "e16c04dc92d72f21e82287fb73b22907f9fcdd35", "parents": [ "0e6e74162164d908edf7889ac66dca09e7505745" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:13:53 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:13:53 2007 -0700" }, "message": "IB/mlx4: Get rid of max_inline_data calculation\n\nThe calculation of max_inline_data in set_kernel_sq_size() is bogus,\nsince it doesn\u0027t take into account the fact that inline segments may\nnot cross a 64-byte boundary, and hence multiple inline segments will\nprobably need to be used to post large inline sends.\n\nWe don\u0027t support inline sends for kernel QPs anyway, so there\u0027s no\npoint in doing this calculation anyway, since the field is just zeroed\nout a little later. So just delete the bogus calculation.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "0e6e74162164d908edf7889ac66dca09e7505745", "tree": "7828d670c099771492dac599f8a7323c4e681e52", "parents": [ "42c059ea2b0aac5f961253ba81c1b464d181a600" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:13:48 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jun 18 08:13:48 2007 -0700" }, "message": "IB/mlx4: Handle new FW requirement for send request prefetching\n\nNew ConnectX firmware introduces FW command interface revision 2,\nwhich requires that for each QP, a chunk of send queue entries (the\n\"headroom\") is kept marked as invalid, so that the HCA doesn\u0027t get\nconfused if it prefetches entries that haven\u0027t been posted yet. Add\ncode to the driver to do this, and also update the user ABI so that\nuserspace can request that the prefetcher be turned off for userspace\nQPs (we just leave the prefetcher on for all kernel QPs).\n\nUnfortunately, marking send queue entries this way is confuses older\nfirmware, so we change the driver to allow only FW command interface\nrevisions 2. This means that users will have to update their firmware\nto work with the new driver, but the firmware is changing quickly and\nthe old firmware has lots of other bugs anyway, so this shouldn\u0027t be too\nbig a deal.\n\nBased on a patch from Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "42c059ea2b0aac5f961253ba81c1b464d181a600", "tree": "44784011ab1fde7212855b07d97926bd6d2a2cae", "parents": [ "614c3c85b5b4c3776439d464939c123cce679dee" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jun 12 10:52:02 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jun 12 10:52:02 2007 -0700" }, "message": "IB/mlx4: Fix warning in rounding up queue sizes\n\nDoing max(1, foo) where foo is u32 generates a warning, because 1 is a\nsigned constant. Fix this by using 1U instead.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "614c3c85b5b4c3776439d464939c123cce679dee", "tree": "9e25ae6a6752eb32c31dfe9d0b55c79b90385fc6", "parents": [ "99f9f3d49cbc7d944476f6fde53a77ec789ab2aa" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jun 12 10:50:42 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jun 12 10:50:42 2007 -0700" }, "message": "IB/mlx4: Fix handling of wq-\u003etail for send completions\n\nCast the increment added to wq-\u003etail when send completions are\nprocessed to u16 to avoid using wrong values caused by standard\ninteger promotions.\n\nThe same bug was fixed in libmlx4 by Eli Cohen \u003celi@mellanox.co.il\u003e.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "99f9f3d49cbc7d944476f6fde53a77ec789ab2aa", "tree": "9b6fdfb39ca8a500e458dbd87dc6e1985d394564", "parents": [ "22353f35c895acb7a8ca27ebdc6397c993b4213e", "a4cd7ed86ff511aebcc97675937039f2321d6987" ], "author": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Mon Jun 11 15:46:08 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Mon Jun 11 15:46:08 2007 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:\n IB/mlx4: Make sure RQ allocation is always valid\n RDMA/cma: Fix initialization of next_port\n IB/mlx4: Fix zeroing of rnr_retry value in ib_modify_qp()\n mlx4_core: Don\u0027t set MTT address in dMPT entries with PA set\n mlx4_core: Check firmware command interface revision\n IB/mthca, mlx4_core: Fix typo in comment\n mlx4_core: Free catastrophic error MSI-X interrupt with correct dev_id\n mlx4_core: Initialize ctx_list and ctx_lock earlier\n mlx4_core: Fix CQ context layout\n" }, { "commit": "a4cd7ed86ff511aebcc97675937039f2321d6987", "tree": "1fdcec754e2086a0f67022c967d55fab12f7aec4", "parents": [ "bf2944bd56c7a48cc3962a860dbc4ceee6b1ace8" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:39 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:39 2007 -0700" }, "message": "IB/mlx4: Make sure RQ allocation is always valid\n\nQPs attached to an SRQ must never have their own RQ, and QPs not\nattached to SRQs must have an RQ with at least 1 entry. Enforce all\nof this in set_rq_size().\n\nBased on a patch by Eli Cohen \u003celi@mellanox.co.il\u003e.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "bf2944bd56c7a48cc3962a860dbc4ceee6b1ace8", "tree": "b6255f3830965ddc615a0b872bd3c4bbfa1bf492", "parents": [ "57f01b53398baebd809e7efd49fc10c10174b46d" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Tue Jun 05 09:57:31 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:38 2007 -0700" }, "message": "RDMA/cma: Fix initialization of next_port\n\nnext_port should be between sysctl_local_port_range[0] and [1].\nHowever, it is initially set to a random value with get_random_bytes(). \nIf the value is negative when treated as a signed integer, next_port\ncan end up outside the expected range because of the result of the % \noperator being negative.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "57f01b53398baebd809e7efd49fc10c10174b46d", "tree": "dc3ab265d42b52f43679cead18d5e35e8a9eba3b", "parents": [ "b2d9308ae43a9cfa56cc9682dc6ed63347a8d2bf" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Wed Jun 06 19:35:04 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:38 2007 -0700" }, "message": "IB/mlx4: Fix zeroing of rnr_retry value in ib_modify_qp()\n\nThe code in __mlx4_ib_modify_qp() overwrites context-\u003eparams1 after\nthe RNR retry parameter is ORed in, which results in the RNR retry\nparameter always being set to 0. Fix this by moving where we OR in\nthe value to later in the function, after the initial assignment of\ncontext-\u003eparams1.\n\nFound by the Mellanox firmware group.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "42f811b8bcdf6695bf74de580b1daf53445e8949", "tree": "c7c9dd321c63eb24c5d41488b47b7178c6c9b572", "parents": [ "8d76527e728d00d1cf9d5dd663caffb2dcf05ae6" ], "author": { "name": "Herbert Xu", "email": "herbert@gondor.apana.org.au", "time": "Mon Jun 04 23:34:44 2007 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@sunset.davemloft.net", "time": "Thu Jun 07 13:39:13 2007 -0700" }, "message": "[IPV4]: Convert IPv4 devconf to an array\n\nThis patch converts the ipv4_devconf config members (everything except\nsysctl) to an array. This allows easier manipulation which will be\nneeded later on to provide better management of default config values.\n\nSigned-off-by: Herbert Xu \u003cherbert@gondor.apana.org.au\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "3e1db334dca97df9034ebeec4657329a2b37a811", "tree": "df9ed4b79daf80e42032fd92e17df448343ee6b3", "parents": [ "2c5cb2355843f9958e19a4b243456be92f97c73b" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sun Jun 03 19:47:10 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 11:51:59 2007 -0700" }, "message": "IB/mthca, mlx4_core: Fix typo in comment\n\ns/signifant/significant/\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "d998ccce020e2cfcf11c6b57503532930ede2894", "tree": "ee9756ab7c9bf087dc47ecd43c1e3a4c439a26c5", "parents": [ "ec56dc0b7f6c3fec20bbc2e98ff1a06edf2fc9b9" ], "author": { "name": "Sean Hefty", "email": "sean.hefty@intel.com", "time": "Mon May 21 17:38:02 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 29 16:07:09 2007 -0700" }, "message": "IB/cm: Fix stale connection detection\n\nThe ib_cm can incorrectly detect a stale connection (a new connection\nrequest for a QPN that is already connected) as a duplicate connection\nrequest. Separate the handling of potential duplicate REQs from stale\nconnections.\n\nSigned-off-by: Sean Hefty \u003csean.hefty@intel.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "ec56dc0b7f6c3fec20bbc2e98ff1a06edf2fc9b9", "tree": "3bb5379d1bc2cd42526fe3d97b728f0d743bca30", "parents": [ "8b7e15772a286d0ef8e4f8eca422ce5368b6fa97" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Mon May 28 14:37:27 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 29 16:07:09 2007 -0700" }, "message": "IPoIB/cm: Fix performance regression on Mellanox\n\ncommit 518b1646 (\"IPoIB/cm: Fix SRQ WR leak\") introduced a severe\nperformance regression on Mellanox cards, because keeping a QP in the\nerror state for extended periods of time moves hardware to the slow\npath (until the QP is destroyed). For example, MPI latency goes from\n~3 usecs to ~7 usecs.\n\nFix this by posting a send WR on one of the QPs that are being\nflushed, instead of using a separate drain QP that is kept in the\nerror state.\n\nThis fixes bug \u003chttps://bugs.openfabrics.org/show_bug.cgi?id\u003d636\u003e,\nreported and bisected by Scott Weitzenkamp at Cisco and debugged by\nSasha Mikheev at Voltaire.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8b7e15772a286d0ef8e4f8eca422ce5368b6fa97", "tree": "646d1d449b0f123d2567a16c5c2a341b17b67933", "parents": [ "6e98ee75c3ab99db48ecc0615c2246dc193111a9" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Sun May 27 18:06:42 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 29 16:07:09 2007 -0700" }, "message": "IB/mthca: Fix handling of send CQE with error for QPs connected to SRQ\n\nmthca_free_err_wqe() currently treats both send and receive CQEs\nidentically if a QP is using an SRQ. But for Tavor hardware, send\nCQEs with error can be chained together even if the RQ is part of SRQ,\nso we may miss some CQEs.\n\nFix by following the WQE chain for all send CQEs even for non-SRQ QPs.\n\nThis fixes crashes in IPoIB CM:\n\u003chttps://bugs.openfabrics.org//show_bug.cgi?id\u003d604\u003e\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "2dfbfc37121d307e1f1d24c2979382cb17b19347", "tree": "b301f2f55738bb6ec9621fbe81161d93701c6bcf", "parents": [ "8fd357a6e3375083f7d321413eb8f6739491f342" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Thu May 24 18:32:46 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu May 24 14:02:40 2007 -0700" }, "message": "IPoIB/cm: Drain cq in ipoib_cm_dev_stop()\n\nSince NAPI polling is disabled while ipoib_cm_dev_stop() is running,\nipoib_cm_dev_stop() must poll the CQ itself in order to see the\npackets draining.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8fd357a6e3375083f7d321413eb8f6739491f342", "tree": "5296052431ba9ca57184c9de827329f6a563f115", "parents": [ "65a2c841d68ae3402ea4cad8d00fe4b9b0a5bc80" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Thu May 24 14:02:39 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu May 24 14:02:39 2007 -0700" }, "message": "IPoIB/cm: Fix timeout check in ipoib_cm_dev_stop()\n\ntime_after() was used backwards, so the timeout occurred immediately.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" } ], "next": "65a2c841d68ae3402ea4cad8d00fe4b9b0a5bc80" }