)]}' { "log": [ { "commit": "b262e60309e1b0eb25d300c7e739427d5316abb1", "tree": "bf319d78c79bb5cb617ff0c8340c73aa349bba15", "parents": [ "93c8b90f01f0dc73891da4e84b26524b61d29d66", "0523820482dcb42784572ffd2296c2f08c275a2b" ], "author": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Wed Oct 01 06:12:56 2008 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Wed Oct 01 06:12:56 2008 -0700" }, "message": "Merge branch \u0027master\u0027 of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6\n\nConflicts:\n\n\tdrivers/net/wireless/ath9k/core.c\n\tdrivers/net/wireless/ath9k/main.c\n\tnet/core/dev.c\n" }, { "commit": "6526128c18564697b301030a7bbf928209e49ae0", "tree": "32cc9e5c50697bc25d1e0a6b7880055c54638211", "parents": [ "c2d1eb085d2eb47a3b16726f79fcf0d53648ebc0" ], "author": { "name": "Andrew Morton", "email": "akpm@linux-foundation.org", "time": "Fri Sep 05 14:04:07 2008 -0700" }, "committer": { "name": "Jeff Garzik", "email": "jgarzik@redhat.com", "time": "Wed Sep 24 18:48:57 2008 -0400" }, "message": "drivers/net/mlx4/alloc.c needs mm.h\n\nsparc32 allmodconfig with linux-next:\n\ndrivers/net/mlx4/alloc.c: In function \u0027mlx4_buf_alloc\u0027:\ndrivers/net/mlx4/alloc.c:164: error: \u0027PAGE_KERNEL\u0027 undeclared (first use in this function)\ndrivers/net/mlx4/alloc.c:164: error: (Each undeclared identifier is reported only once\ndrivers/net/mlx4/alloc.c:164: error: for each function it appears in.)\n\nthis is due to some header shuffle in linux-next. I didn\u0027t look to see what\nit was. I\u0027d sugges that this patch be merged ahead of a linux-next merge to\navoid bisection breaks.\n\nWe strictly only need asm/pgtable.h, but going direct to asm includes always\nseems grubby.\n\nCc: Jeff Garzik \u003cjeff@garzik.org\u003e\nCc: \"David S. Miller\" \u003cdavem@davemloft.net\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Jeff Garzik \u003cjgarzik@redhat.com\u003e\n" }, { "commit": "29bdc88384c2b24e37e5760df0dc898546083d6b", "tree": "9d8fe16ac4728ab669f81a763effda21301d9ddf", "parents": [ "c9257433f2eaf8803a1f3d3be5d984232db41ffe" ], "author": { "name": "Vladimir Sokolovsky", "email": "vlad@mellanox.co.il", "time": "Mon Sep 15 14:25:23 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Sep 15 14:25:23 2008 -0700" }, "message": "IB/mlx4: Fix up fast register page list format\n\nByte swap the addresses in the page list for fast register work requests\nto big endian to match what the HCA expectx. Also, the addresses must\nhave the \"present\" bit set so that the HCA knows it can access them.\nOtherwise the HCA will fault the first time it accesses the memory\nregion.\n\nSigned-off-by: Vladimir Sokolovsky \u003cvlad@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "c9257433f2eaf8803a1f3d3be5d984232db41ffe", "tree": "20d09185709dee8aad0f9bb7738724184b3a1f14", "parents": [ "1941246dd98089dd637f44d3bd4f6cc1c61aa9e4" ], "author": { "name": "Vladimir Sokolovsky", "email": "vlad@mellanox.co.il", "time": "Tue Sep 02 13:38:29 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Sep 02 13:38:29 2008 -0700" }, "message": "mlx4_core: Set RAE and init mtt_sz field in FRMR MPT entries\n\nSet the RAE (remote access enable) bit and correctly initialize the\nMTT size in MPT entries being set up for fast register memory\nregions. Otherwise the callers can\u0027t enable remote access and in fact\ncan\u0027t fast register at all (since the HCA will think no MTT entries\nare allocated).\n\nSigned-off-by: Vladimir Sokolovsky \u003cvlad@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8be1a6d6c77ab4532e4476fdb8177030ef48b52c", "tree": "c94ff327a6291672e1f9071a79bf61d63f5f0cb9", "parents": [ "852fef69c0d9510a28a70221cfddd004efa02552", "cc9969c96765476db5e30c9e074a57838604c1d1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 26 20:40:36 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 26 20:40:36 2008 -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 mlx4: Update/add Mellanox Technologies copyright lines to mlx4 driver files\n mlx4_core: Add VLAN tag field to WQE control segment struct\n RDMA/nes: CM connection setup/teardown rework\n IPoIB: Correct help text for INFINIBAND_IPOIB_DEBUG\n IPoIB/cm: Connected mode is no longer EXPERIMENTAL\n RDMA/ucm: BKL is not needed for ib_ucm_open()\n RDMA/ucma: BKL is not needed for ucma_open()\n" }, { "commit": "8d8bb39b9eba32dd70e87fd5ad5c5dd4ba118e06", "tree": "64090a84f4c4466f9f30ff46c993e0cede379052", "parents": [ "c485b465a031b6f9b9a51300e0ee1f86efc6db87" ], "author": { "name": "FUJITA Tomonori", "email": "fujita.tomonori@lab.ntt.co.jp", "time": "Fri Jul 25 19:44:49 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 26 12:00:03 2008 -0700" }, "message": "dma-mapping: add the device argument to dma_mapping_error()\n\nAdd per-device dma_mapping_ops support for CONFIG_X86_64 as POWER\narchitecture does:\n\nThis enables us to cleanly fix the Calgary IOMMU issue that some devices\nare not behind the IOMMU (http://lkml.org/lkml/2008/5/8/423).\n\nI think that per-device dma_mapping_ops support would be also helpful for\nKVM people to support PCI passthrough but Andi thinks that this makes it\ndifficult to support the PCI passthrough (see the above thread). So I\nCC\u0027ed this to KVM camp. Comments are appreciated.\n\nA pointer to dma_mapping_ops to struct dev_archdata is added. If the\npointer is non NULL, DMA operations in asm/dma-mapping.h use it. If it\u0027s\nNULL, the system-wide dma_ops pointer is used as before.\n\nIf it\u0027s useful for KVM people, I plan to implement a mechanism to register\na hook called when a new pci (or dma capable) device is created (it works\nwith hot plugging). It enables IOMMUs to set up an appropriate\ndma_mapping_ops per device.\n\nThe major obstacle is that dma_mapping_error doesn\u0027t take a pointer to the\ndevice unlike other DMA operations. So x86 can\u0027t have dma_mapping_ops per\ndevice. Note all the POWER IOMMUs use the same dma_mapping_error function\nso this is not a problem for POWER but x86 IOMMUs use different\ndma_mapping_error functions.\n\nThe first patch adds the device argument to dma_mapping_error. The patch\nis trivial but large since it touches lots of drivers and dma-mapping.h in\nall the architecture.\n\nThis patch:\n\ndma_mapping_error() doesn\u0027t take a pointer to the device unlike other DMA\noperations. So we can\u0027t have dma_mapping_ops per device.\n\nNote that POWER already has dma_mapping_ops per device but all the POWER\nIOMMUs use the same dma_mapping_error function. x86 IOMMUs use device\nargument.\n\n[akpm@linux-foundation.org: fix sge]\n[akpm@linux-foundation.org: fix svc_rdma]\n[akpm@linux-foundation.org: build fix]\n[akpm@linux-foundation.org: fix bnx2x]\n[akpm@linux-foundation.org: fix s2io]\n[akpm@linux-foundation.org: fix pasemi_mac]\n[akpm@linux-foundation.org: fix sdhci]\n[akpm@linux-foundation.org: build fix]\n[akpm@linux-foundation.org: fix sparc]\n[akpm@linux-foundation.org: fix ibmvscsi]\nSigned-off-by: FUJITA Tomonori \u003cfujita.tomonori@lab.ntt.co.jp\u003e\nCc: Muli Ben-Yehuda \u003cmuli@il.ibm.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nCc: Avi Kivity \u003cavi@qumranet.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "51a379d0c8f7a6db7c9e3c9c770d90a6d2d1ef9b", "tree": "c3f2aecdaaa8e528556f7eeeffd71b983a38bbbc", "parents": [ "25c94d010a8ae8605dc4d5453e0c82fa97da5d12" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Fri Jul 25 10:32:52 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Fri Jul 25 10:32:52 2008 -0700" }, "message": "mlx4: Update/add Mellanox Technologies copyright lines to mlx4 driver files\n\nUpdate existing Mellanox copyright lines to 2008, and add such lines\nto files where they are missing.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "5c402355adf8f920531f02099f4ec0d2bccd4c64", "tree": "eba326f8df01efc0d384874839384040401f5b45", "parents": [ "ecc8b655b38a880b578146895e0e1e2d477ca2c0", "2cc177364e4746becdf421f926fb967c047ccc32" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jul 24 12:56:07 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jul 24 12:56:07 2008 -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 MAINTAINERS: Remove Glenn Streiff from NetEffect entry\n mlx4_core: Improve error message when not enough UAR pages are available\n IB/mlx4: Add support for memory management extensions and local DMA L_Key\n IB/mthca: Keep free count for MTT buddy allocator\n mlx4_core: Keep free count for MTT buddy allocator\n mlx4_code: Add missing FW status return code\n IB/mlx4: Rename struct mlx4_lso_seg to mlx4_wqe_lso_seg\n mlx4_core: Add module parameter to enable QoS support\n RDMA/iwcm: Remove IB_ACCESS_LOCAL_WRITE from remote QP attributes\n IPoIB: Include err code in trace message for ib_sa_path_rec_get() failures\n IB/sa_query: Check if sm_ah is NULL in ib_sa_remove_one()\n IB/ehca: Release mutex in error path of alloc_small_queue_page()\n IB/ehca: Use default value for Local CA ACK Delay if FW returns 0\n IB/ehca: Filter PATH_MIG events if QP was never armed\n IB/iser: Add support for RDMA_CM_EVENT_ADDR_CHANGE event\n RDMA/cma: Add RDMA_CM_EVENT_TIMEWAIT_EXIT event\n RDMA/cma: Add RDMA_CM_EVENT_ADDR_CHANGE event\n" }, { "commit": "27ac792ca0b0a1e7e65f20342260650516c95864", "tree": "8e0bc93612da0803fe12303ccb75c837cd633c83", "parents": [ "d92bc318547507a944a22e7ef936793dc0fe167f" ], "author": { "name": "Andrea Righi", "email": "righi.andrea@gmail.com", "time": "Wed Jul 23 21:28:13 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jul 24 10:47:21 2008 -0700" }, "message": "PAGE_ALIGN(): correctly handle 64-bit values on 32-bit architectures\n\nOn 32-bit architectures PAGE_ALIGN() truncates 64-bit values to the 32-bit\nboundary. For example:\n\n\tu64 val \u003d PAGE_ALIGN(size);\n\nalways returns a value \u003c 4GB even if size is greater than 4GB.\n\nThe problem resides in PAGE_MASK definition (from include/asm-x86/page.h for\nexample):\n\n#define PAGE_SHIFT 12\n#define PAGE_SIZE (_AC(1,UL) \u003c\u003c PAGE_SHIFT)\n#define PAGE_MASK (~(PAGE_SIZE-1))\n...\n#define PAGE_ALIGN(addr) (((addr)+PAGE_SIZE-1)\u0026PAGE_MASK)\n\nThe \"~\" is performed on a 32-bit value, so everything in \"and\" with\nPAGE_MASK greater than 4GB will be truncated to the 32-bit boundary.\nUsing the ALIGN() macro seems to be the right way, because it uses\ntypeof(addr) for the mask.\n\nAlso move the PAGE_ALIGN() definitions out of include/asm-*/page.h in\ninclude/linux/mm.h.\n\nSee also lkml discussion: http://lkml.org/lkml/2008/6/11/237\n\n[akpm@linux-foundation.org: fix drivers/media/video/uvc/uvc_queue.c]\n[akpm@linux-foundation.org: fix v850]\n[akpm@linux-foundation.org: fix powerpc]\n[akpm@linux-foundation.org: fix arm]\n[akpm@linux-foundation.org: fix mips]\n[akpm@linux-foundation.org: fix drivers/media/video/pvrusb2/pvrusb2-dvb.c]\n[akpm@linux-foundation.org: fix drivers/mtd/maps/uclinux.c]\n[akpm@linux-foundation.org: fix powerpc]\nSigned-off-by: Andrea Righi \u003crighi.andrea@gmail.com\u003e\nCc: \u003clinux-arch@vger.kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "76442640829163d0cdb67c2bf0cb4b81a0fe537b", "tree": "8c5cd9094d24a8181d64cd840636ab9885c8e1c1", "parents": [ "95d04f0735b4fc837bff9aedcc3f3efb20ddc3d1" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Jul 23 08:12:47 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Jul 23 08:12:47 2008 -0700" }, "message": "mlx4_core: Improve error message when not enough UAR pages are available\n\nIf an mlx4 device with default FW (which gives a UAR BAR size of 8 MB)\nis used in a system with 64 KB pages, then there are only 8192/64\u003d\u003d128\nUAR pages available. However, the first 128 UAR pages are reserved\nfor use with event queue doorbells, so no UAR pages are available to\ndo anything else with, which means that the driver cannot work.\n\nThe current driver fails with a fairly cryptic \"Failed to allocate\ndriver access region, aborting\" message in this situation. Fix the\ndriver to detect the problem earlier and print out a clearer\ndescription of the problem and a suggestion of how to fix it (use a\nnew firmware image).\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "95d04f0735b4fc837bff9aedcc3f3efb20ddc3d1", "tree": "900cd7debae9827c3e20f5199307ae27e83ba862", "parents": [ "e4044cfc493338cd09870bd45dc646336bb66e9f" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Jul 23 08:12:26 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Jul 23 08:12:26 2008 -0700" }, "message": "IB/mlx4: Add support for memory management extensions and local DMA L_Key\n\nAdd support for the following operations to mlx4 when device firmware\nsupports them:\n\n - Send with invalidate and local invalidate send queue work requests;\n - Allocate/free fast register MRs;\n - Allocate/free fast register MR page lists;\n - Fast register MR send queue work requests;\n - Local DMA L_Key.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e4044cfc493338cd09870bd45dc646336bb66e9f", "tree": "2385c18ee6a79393b4e1c5c25ee1f0f980d7da93", "parents": [ "899698dad72340b562478b8b770317f2f0fe0c09" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 22 14:19:40 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 22 14:19:40 2008 -0700" }, "message": "mlx4_core: Keep free count for MTT buddy allocator\n\nMTT entries are allocated with a buddy allocator, which just keeps\nbitmaps for each level of the buddy table. However, all free space\nstarts out at the highest order, and small allocations start scanning\nfrom the lowest order. When the lowest order tables have no free\nspace, this can lead to scanning potentially millions of bits before\nfinding a free entry at a higher order.\n\nWe can avoid this by just keeping a count of how many free entries\neach order has, and skipping the bitmap scan when an order is\ncompletely empty. This provides a nice performance boost for a\nnegligible increase in memory usage.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "899698dad72340b562478b8b770317f2f0fe0c09", "tree": "117856054c5b6c41b3be86c43846abff004d9f39", "parents": [ "47b374752aed1c029f995473c7c463ee3ae5fbaa" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Tue Jul 22 14:19:39 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 22 14:19:39 2008 -0700" }, "message": "mlx4_code: Add missing FW status return code\n\nAdd ICM_ERROR firmware status code. In mapping to errnos, -ENFILE\nseems closest.\n\nThis is in preparation for providing more detailed log info using\nmlx4_err() in low-level driver when a non-zero status is returned.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "51f5f0ee22b98980f7816d42647467cd5f4b3b45", "tree": "0876e9302e9183293502395ee57471a8411b9109", "parents": [ "d35cb360c29956510b2fe1a953bd4968536f7216" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Tue Jul 22 14:19:37 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 22 14:19:37 2008 -0700" }, "message": "mlx4_core: Add module parameter to enable QoS support\n\nAdd a module parameter \"enable_qos\" to mlx4_core. If this param is\nset, enable support for QoS in the INIT_HCA command. By default, the\nparameter is set to 0 (disabled).\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "2d92865158d0e21ef4350703af64bc2a610d81d3", "tree": "8b11b29982ccff35bc6810e5a57c568ce1a4ada0", "parents": [ "de910bd92137005b5e1ecaf2ce68053d7d7d5350" ], "author": { "name": "Vladimir Sokolovsky", "email": "vlad@mellanox.co.il", "time": "Mon Jul 14 23:48:53 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 14 23:48:53 2008 -0700" }, "message": "mlx4_core: Use MOD_STAT_CFG command to get minimal page size\n\nThere was a bug in some versions of the mlx4 driver in\nmlx4_alloc_fmr(), which hardcoded the minimum acceptable page_shift to\nbe 12. However, new ConnectX firmware can support a minimum\npage_shift of 9 (log_pg_sz of 9 returned by QUERY_DEV_LIM) -- so with\nold drivers, ib_fmr_alloc() would fail for ULPs using the device\nminimum when creating FMRs.\n\nTo preserve firmware compatibility with released mlx4 drivers, the\nfirmware will continue to return 12 as before for log_page_sz in\nQUERY_DEV_CAP for these drivers. However, to enable new drivers to\ntake advantage of the available smaller page size, the mlx4 driver now\nfirst sets the log_pg_sz to the device minimum by setting a\nlog_page_sz value to 0 via the MOD_STAT_CFG command and then reading\nthe real minimum via QUERY_DEV_CAP.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@mellanox.co.il\u003e\nSigned-off-by: Vladimir Sokolovsky \u003cvlad@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "521e575b9a7324a0bca762622139f69582a042bf", "tree": "5962695835c6d11b424d76c64421ebd436b52a34", "parents": [ "47ee1b9f2e7bf73950602efe0b74fa1a8481f222" ], "author": { "name": "Ron Livne", "email": "ronli@voltaire.com", "time": "Mon Jul 14 23:48:48 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 14 23:48:48 2008 -0700" }, "message": "IB/mlx4: Add support for blocking multicast loopback packets\n\nAdd support for handling the IB_QP_CREATE_MULTICAST_BLOCK_LOOPBACK\nflag by using the per-multicast group loopback blocking feature of\nmlx4 hardware.\n\nSigned-off-by: Ron Livne \u003cronli@voltaire.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "c5057ddccbcb4bf363af628d7963a7475f4114a7", "tree": "4f304c61335dc607d51e6b242d058c430388708e", "parents": [ "cf04690885972eaba830ee761de545a6956197e6" ], "author": { "name": "Oren Duer", "email": "oren@mellanox.co.il", "time": "Mon May 05 15:56:52 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon May 05 15:56:52 2008 -0700" }, "message": "mlx4_core: Support creation of FMRs with pages smaller than 4K\n\nDon\u0027t hard code a test against a minimum page shift of 12, since the\ndevice may support smaller pages. Test against the actual smallest\npage size from the device capabilities.\n\nSigned-off-by: Oren Duer \u003coren@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": "bbdc2821db041fb07ffa52e4a0e1ebb5410790e9", "tree": "1a0757409c89c6236d9c0ffa1eb0f5a63faa2527", "parents": [ "d227fa7288adebe5ba37fa8e4a589c977d4e4a34" ], "author": { "name": "Olaf Kirch", "email": "okir@lst.de", "time": "Tue Apr 29 13:46:53 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Apr 29 13:46:53 2008 -0700" }, "message": "mlx4_core: Avoid recycling old FMR R_Keys too soon\n\nWhen a FMR is unmapped, mlx4 resets the map count to 0, and clears the\nupper part of the R_Key which is used as the sequence counter.\n\nThis poses a problem for RDS, which uses ib_fmr_unmap as a fence\noperation. RDS assumes that after issuing an unmap, the old R_Keys\nwill be invalid for a \"reasonable\" period of time. For instance,\nOracle processes uses shared memory buffers allocated from a pool of\nbuffers. When a process dies, we want to reclaim these buffers -- but\nwe must make sure there are no pending RDMA operations to/from those\nbuffers. The only way to achieve that is by using unmap and sync the\nTPT.\n\nHowever, when the sequence count is reset on unmap, there is a high\nlikelihood that a new mapping will be given the same R_Key that was\nissued a few milliseconds ago.\n\nTo prevent this, don\u0027t reset the sequence count when unmapping a FMR.\n\nSigned-off-by: Olaf Kirch \u003colaf.kirch@oracle.com\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e463c7b197dbe64b8a99b0612c65f286937e5bf1", "tree": "aadd360d04b9cdf7eaf7d4eed1f79bd29a3b12df", "parents": [ "7663c1e2792a9662b23dec6e19bfcd3d55360b8f" ], "author": { "name": "Yevgeny Petrilin", "email": "yevgenyp@mellanox.co.il", "time": "Tue Apr 29 13:46:50 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Apr 29 13:46:50 2008 -0700" }, "message": "mlx4_core: Add a way to set the \"collapsed\" CQ flag\n\nExtend the mlx4_cq_resize() API with a way to set the \"collapsed\" flag\nfor the CQ being created.\n\nSigned-off-by: Yevgeny Petrilin \u003cyevgenyp@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "ed4d3c1061d6f367a4ef5e1656c25af3314fe2b7", "tree": "93e627fa3d8f7b439625ffec5ccd390d100d67a1", "parents": [ "38ae6a535470b959df67ded6798fc542bb212e19" ], "author": { "name": "Yevgeny Petrilin", "email": "yevgenyp@mellanox.co.il", "time": "Fri Apr 25 14:52:32 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Fri Apr 25 14:52:32 2008 -0700" }, "message": "mlx4_core: Add helper to move QP to ready-to-send\n\nAvoid duplicating code in ethernet and FC modules.\n\nSigned-off-by: Yevgeny Petrilin \u003cyevgenyp@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "38ae6a535470b959df67ded6798fc542bb212e19", "tree": "35ee1d5450f0b0f491764ebd22babceb9cbf8f1d", "parents": [ "31d1e340f0e8d53804d737571b2f2bb28a74ecc5" ], "author": { "name": "Yevgeny Petrilin", "email": "yevgenyp@mellanox.co.il", "time": "Fri Apr 25 14:27:08 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Fri Apr 25 14:27:08 2008 -0700" }, "message": "mlx4_core: Add HW queues allocation helpers\n\nWrap doorbell, buffer and MTT allocation in helper functions for\nethernet and FC modules to use.\n\nSigned-off-by: Yevgeny Petrilin \u003cyevgenyp@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "f5b3a096b138940f283907debe9bde6c6f40ebf3", "tree": "a999540103ef6f7c4d1be52689c326518d4dabe5", "parents": [ "6296883ca4cd52dafb45f191d24102e28ded38f2" ], "author": { "name": "Vladimir Sokolovsky", "email": "vlad@mellanox.co.il", "time": "Wed Apr 23 11:55:45 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 23 11:55:45 2008 -0700" }, "message": "mlx4_core: CQ resizing should pass a 0 opcode modifier to MODIFY_CQ\n\nThe call to mlx4_MODIFY_CQ() had a typo so that mlx4_cq_resize() was\nactually asking the FW to modify a CQ\u0027s interrupt moderation rather than\nasking it to resize a CQ.\n\nSigned-off-by: Vladimir Sokolovsky \u003cvlad@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "6296883ca4cd52dafb45f191d24102e28ded38f2", "tree": "341e90a9560d8cf6b498d249a6ac81aeea97dd7b", "parents": [ "14fb05b3497351fbeb514381bcd227d84e115bd9" ], "author": { "name": "Yevgeny Petrilin", "email": "yevgenyp@mellanox.co.il", "time": "Wed Apr 23 11:55:45 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 23 11:55:45 2008 -0700" }, "message": "mlx4_core: Move kernel doorbell management into core\n\nIn addition to mlx4_ib, there will be ethernet and FC consumers of\nmlx4_core, so move the code for managing kernel doorbells into the\ncore module to avoid having to duplicate this multiple times.\n\nSigned-off-by: Yevgeny Petrilin \u003cyevgenyp@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "4ff08a76bce3ed5f7d109c646dcfc474d7946213", "tree": "9b0f8cd225c79ac8acf65125bd1f6053d6bd6425", "parents": [ "4dc51b32582d45cb7e8322d750ffe0e6d82b506d" ], "author": { "name": "Eli Cohen", "email": "eli@dev.mellanox.co.il", "time": "Wed Apr 16 21:09:35 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:09:35 2008 -0700" }, "message": "IB/mlx4: Fix incorrect comment\n\nmlx4 hardware does not support external DDR memory. Moreover, UAR\narea (BAR 2) can change depending on FW version.\n\nSigned-off-by: Eli Cohen \u003celi@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "4dc51b32582d45cb7e8322d750ffe0e6d82b506d", "tree": "0d8549a6cc82b9719086a2673c76de8cd5b161e9", "parents": [ "c83b5b1cb2b050c7a9054f330598df99c31abb98" ], "author": { "name": "Eli Cohen", "email": "eli@dev.mellanox.co.il", "time": "Wed Apr 16 21:09:35 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:09:35 2008 -0700" }, "message": "IB/mlx4: Fix race when detaching a QP from a multicast group\n\nWhen detaching the last QP from an MCG entry, we need to make\nsure that at any time, there will be no entry with zero number of\nQPs which is linked to the list of the MCGs of the corresponding\nhash index. So don\u0027t write back the MCG entry if we are removing the\nlast QP; just unlink the entry.\n\nAlso, remove an unnecessary MCG read when attaching a QP requires\nallocation of a new entry in the AMGM.\n\nSigned-off-by: Eli Cohen \u003celi@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "bbf8eed1a0f8949f7385146624f736f829992a70", "tree": "5cf6a5c76ca4c038d3ca0f53abc5f2976872696b", "parents": [ "3fdcb97f0b8d8a29117dc36acd0b15965d2a2160" ], "author": { "name": "Vladimir Sokolovsky", "email": "vlad@dev.mellanox.co.il", "time": "Wed Apr 16 21:09:33 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:09:33 2008 -0700" }, "message": "IB/mlx4: Add support for resizing CQs\n\nSigned-off-by: Vladimir Sokolovsky \u003cvlad@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "3fdcb97f0b8d8a29117dc36acd0b15965d2a2160", "tree": "a8bad6e48e9654f10e1b8ebfde3b086d83d2756e", "parents": [ "28d52b3cd8d48ef0ff77d4a8a7a21fc2816bb0a5" ], "author": { "name": "Eli Cohen", "email": "eli@dev.mellanox.co.il", "time": "Wed Apr 16 21:09:33 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:09:33 2008 -0700" }, "message": "IB/mlx4: Add support for modifying CQ moderation parameters\n\nSigned-off-by: Eli Cohen \u003celi@mellnaox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "9b1f38515c41a5f13021e15b50b6558db236d6d5", "tree": "80e491019a67cb3337b34cde4628876e3c94553d", "parents": [ "139b2db7951ed0c808884cfebca02a884aebe0c3" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Wed Apr 16 21:09:32 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:09:32 2008 -0700" }, "message": "mlx4_core: Increase max number of QPs to 128K\n\nWith the advent large clusters which utilize multicore hosts, 64K QPs\nis not enough. We should increase the default maximum for QPs to 128K.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b832be1e4007f4a54954ec68bd865ff05d6babca", "tree": "f8780fb17293a5b02cd21fed468e1270daac91d8", "parents": [ "40ca1988e03c001747d0b4cc1b25cf38297c9f9e" ], "author": { "name": "Eli Cohen", "email": "eli@dev.mellanox.co.il", "time": "Wed Apr 16 21:09:27 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:09:27 2008 -0700" }, "message": "IB/mlx4: Add IPoIB LSO support\n\nAdd TSO support to the mlx4_ib driver.\n\nSigned-off-by: Eli Cohen \u003celi@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8ff095ec4bce7be943beff3b330562e2f0e42167", "tree": "d0643b575033a033a072c2c0687be7e57d183d99", "parents": [ "6046136c742e32d5e6431cdcd8957638d1816821" ], "author": { "name": "Eli Cohen", "email": "eli@dev.mellanox.co.il", "time": "Wed Apr 16 21:01:10 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:01:10 2008 -0700" }, "message": "IB/mlx4: Add IPoIB checksum offload support\n\nConnectX devices support checksum generation and verification of TCP\nand UDP packets for UD IPoIB messages. This patch checks if the HCA\nsupports this and sets the IB_DEVICE_UD_IP_CSUM capability flag if it\ndoes. It implements support for handling the IB_SEND_IP_CSUM send\nflag and setting the csum_ok field in receive work completions.\n\nSigned-off-by: Eli Cohen \u003celi@mellanox.co.il\u003e\nSigned-off-by: Ali Ayub \u003cali@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "37608eea86a358ac6a18df0af55d4f77d08a1f30", "tree": "bb0d7c9a93763e6b3fda744a08b2a0c6ed80dfec", "parents": [ "26c4fc26d0af9b16a6a234318d15ee0b3896a63d" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:01:08 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:01:08 2008 -0700" }, "message": "mlx4_core: Fix confusion between mlx4_event and mlx4_dev_event enums\n\nThe struct mlx4_interface.event() method was supposed to get an enum\nmlx4_dev_event, but the driver code was actually passing in the\nhardware enum mlx4_event values. Fix up the callers of\nmlx4_dispatch_event() so that they pass in the right type of value,\nand fix up the event method in mlx4_ib so that it can handle the enum\nmlx4_dev_event values.\n\nThis eliminates the need for the subtype parameter to the event\nmethod, so remove it.\n\nThis also fixes the sparse warning\n\n drivers/net/mlx4/intf.c:127:48: warning: mixing different enum types\n drivers/net/mlx4/intf.c:127:48: int enum mlx4_event versus\n drivers/net/mlx4/intf.c:127:48: int enum mlx4_dev_event\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "ca2812111433284024386a8cfd4151651469523e", "tree": "a4b9e4d4612f0ec8b524e86f74b44bfc4214d9d5", "parents": [ "5d5e815db9651909e435918e46c84dfcbc36154f" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:01:04 2008 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Apr 16 21:01:04 2008 -0700" }, "message": "mlx4_core: Move opening brace of function onto a new line\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "11e75a7455a7bc73e752c0c985986c2b1f8c930a", "tree": "6d3b43ccaeb1b6fc2c8e91f25874baa2ec224bdb", "parents": [ "e6028c0e004d334bb9ed75d4c918f4c763af1b9f" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Feb 14 13:41:29 2008 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Feb 14 10:43:48 2008 -0800" }, "message": "mlx4_core: Move table_find from fmr_alloc to fmr_enable\n\nmlx4_table_find (for FMR MPTs) requires that ICM memory already be\nmapped. Before this fix, FMR allocation depended on ICM memory\nalready being mapped for the MPT entry. If all currently mapped\nentries are taken, the find operation fails (even if the MPT ICM table\nstill had more entries, which were just not mapped yet).\n\nThis fix moves the mpt find operation to fmr_enable, to guarantee that\nany required ICM memory mapping has already occurred.\n\nFound by Oren Duer 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": "29c271123dc7895a9f77d3e61e747b2a052d0a2a", "tree": "0b416b62870d244261f3bb274ccc328ab30f79da", "parents": [ "cc13e442952a347f7e217eeaee4778485394f1be" ], "author": { "name": "Olof Johansson", "email": "olof@lixom.net", "time": "Sun Feb 10 20:22:57 2008 -0600" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Feb 11 14:19:42 2008 -0800" }, "message": "mlx4_core: Fix build break (missing include)\n\nCommit 313abe55 (\"mlx4_core: For 64-bit systems, vmap() kernel queue\nbuffers\") caused this to pop up on powerpc allyesconfig, looks like a\nmissing include file:\n\n drivers/net/mlx4/alloc.c: In function \u0027mlx4_buf_alloc\u0027:\n drivers/net/mlx4/alloc.c:162: error: implicit declaration of function \u0027vmap\u0027\n drivers/net/mlx4/alloc.c:162: error: \u0027VM_MAP\u0027 undeclared (first use in this function)\n drivers/net/mlx4/alloc.c:162: error: (Each undeclared identifier is reported only once\n drivers/net/mlx4/alloc.c:162: error: for each function it appears in.)\n drivers/net/mlx4/alloc.c:162: warning: assignment makes pointer from integer without a cast\n drivers/net/mlx4/alloc.c: In function \u0027mlx4_buf_free\u0027:\n drivers/net/mlx4/alloc.c:187: error: implicit declaration of function \u0027vunmap\u0027\n\nSigned-off-by: Olof Johansson \u003colof@lixom.net\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b57aacfa7a95328f469d0360e49289b023c47e9e", "tree": "5642416e92540b93ffb1221ae75b3f1f43ffcd7b", "parents": [ "313abe55a87bc10e55d00f337d609e17ad5f8c9a" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Feb 06 21:17:59 2008 -0800" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Feb 06 21:17:59 2008 -0800" }, "message": "mlx4_core: Clean up struct mlx4_buf\n\nNow that struct mlx4_buf.u is a struct instead of a union because of\nthe vmap() changes, there\u0027s no point in having a struct at all. So\nmove .direct and .page_list directly into struct mlx4_buf and get rid\nof a bunch of unnecessary \".u\"s.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "313abe55a87bc10e55d00f337d609e17ad5f8c9a", "tree": "833ff3c4b33f83d4ca64ed322c2d8efa21529d71", "parents": [ "1c69fc2a9012e160c8d459f63df74a6b01db8322" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Mon Jan 28 10:40:51 2008 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Feb 06 21:17:45 2008 -0800" }, "message": "mlx4_core: For 64-bit systems, vmap() kernel queue buffers\n\nSince kernel virtual memory is not a problem on 64-bit systems, there\nis no reason to use our own 2-layer page mapping scheme for large\nkernel queue buffers on such systems. Instead, map the page list to a\nsingle virtually contiguous buffer with vmap(), so that can we access\nbuffer memory via direct indexing.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.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": "f33afc26dc03e6e0513e2e300f2aa0ad5463c2d2", "tree": "8d4645dd7c20c466a1c421a2ed05d480497f629f", "parents": [ "68f3948dab39249d178eb007c071f87fb6481fc6" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Feb 04 20:20:44 2008 -0800" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Feb 04 20:20:44 2008 -0800" }, "message": "IB: Avoid marking __devinitdata as const\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "893da75956ab48545e8732b46e1cf4350bd25f9c", "tree": "fe5125ffd8ec7d84f3b4937f6ade9ee65f7a4913", "parents": [ "6ccef1de2c1718729dd1c7ee8bd98473519eb3b3" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Sun Jan 27 18:13:25 2008 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Feb 04 20:20:43 2008 -0800" }, "message": "mlx4_core: Don\u0027t read reserved fields in mlx4_QUERY_ADAPTER()\n\nThe firmware QUERY_ADAPTER command does not return vendor_id,\ndevice_id, and revision_id; eliminate these fields from the query.\n\nInitialize the rev_id field of the mlx4 device via init_node_data (MAD\nIFC query), as is done in the query_device verb implementation.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e8f9b2ed9882874ca96716597bd8c7113289e77b", "tree": "ab18d283baaf14da6ce4a387affd1a24f3d23336", "parents": [ "9ef9dc69d4167276c04590d67ee55de8380bc1ad" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Feb 04 20:20:41 2008 -0800" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Feb 04 20:20:41 2008 -0800" }, "message": "mlx4_core: Fix more section mismatches\n \nCommit 3d73c288 (\"mlx4_core: Fix section mismatches\") fixed some of\nthe section mismatches introduced when error recovery was added, but\nthere were still more cases of errory recovery code calling into\n__devinit code from regular .text. Fix this by getting rid of the\nnow-incorrect __devinit annotations.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "5920869f1ec263fee4255ec71d333c118406a419", "tree": "f9d4c2739feb0c110a65a690b09f323ef5212221", "parents": [ "3c8450860ba9d6279dbc969633eacf99161860d9" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Mon Dec 10 05:25:23 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Fri Jan 25 14:15:29 2008 -0800" }, "message": "mlx4_core: Fix max_eqs masking in QUERY_DEV_CAP\n\nlog_max_eqs is a 4-bit field, not a 3-bit field in the response to the\nQUERY_DEV_CAP FW command, so we should mask with 0xf instead of 0x7\nwhen reading it.\n\nFound by Yossi Leybovitch 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": "9ed87fd34c97a998e63505718ce7e107a23c84c3", "tree": "137e38d8fed9a1b43f34f0ed009ff6ea048ac884", "parents": [ "4187b915a0f7eaa69707715e80d9fc253ff6167a" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Tue Nov 20 13:01:28 2007 -0800" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Nov 20 13:01:28 2007 -0800" }, "message": "mlx4_core: Fix state check in mlx4_qp_modify()\n\nWhen checking the states passed in, mlx4_qp_modify() accidentally checks\ncur_state twice rather than checking cur_state and new_state. Fix this\nto make sure that both values are in-bounds.\n\nSince these values may be passed in from userspace, this bug results in\nuserspace being able to trigger an oops.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nCc: stable \u003cstable@kernel.org\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e383d19e90cfbbf8e00512d44194ce175b3f60a2", "tree": "fb1c2b81505605998cca9062d10b816784f6559f", "parents": [ "9a7666494bac60b99d2bd7d904bd22e8c9b1e3f7" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Wed Nov 14 16:33:27 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Nov 14 08:20:03 2007 -0800" }, "message": "mlx4_core: Fix thinko in QP destroy (incorrect bitmap_free)\n\nFix thinko in commit eaf559bf (\"mlx4_core: Don\u0027t free special QPs in\nQP number bitmap\"). The old commit had the logic exactly backwards\nand ended up freeing *only* special QPs, which not only left the\noriginal bug in place but also introduced the problem that the QP\nnumber bitmap would get full after a while.\n\nFound by Dotan Barak 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": "3bba11e5c47dfc1d381a1ece26464fb7eea2d79c", "tree": "70eab30ede7905899420a52a350eb7ba15f82955", "parents": [ "9418d5dc9ba40b88737580457bf3b7c63c60ec43" ], "author": { "name": "Ali Ayoub", "email": "ali@mellanox.co.il", "time": "Tue Nov 13 15:26:57 2007 -0800" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Nov 13 15:26:57 2007 -0800" }, "message": "mlx4_core: Fix possible bad free in mlx4_buf_free()\n\nWhen mlx4_buf_free() is called from the error path of\nmlx4_buf_alloc(), it may be passed a buffer structure that does not\nhave all pages filled in. Add a check for NULL to mlx4_buf_free() so\nwe avoid passing NULL to dma_free_coherent() (which will crash).\n\nSigned-off-by: Ali Ayoub \u003cali@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": "642f149031d70415d9318b919d50b71e4724adbd", "tree": "e792ad29dedffc6756d55e9d63e18ada35515b4b", "parents": [ "bd6dee6f30a0f6943df190b387b5f8fe98a848f3" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Wed Oct 24 11:20:47 2007 +0200" }, "committer": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Wed Oct 24 11:20:47 2007 +0200" }, "message": "SG: Change sg_set_page() to take length and offset argument\n\nMost drivers need to set length and offset as well, so may as well fold\nthose three lines into one.\n\nAdd sg_assign_page() for those two locations that only needed to set\nthe page, where the offset/length is set outside of the function context.\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "0b776eb5426752d4e53354ac89e3710d857e09a7", "tree": "1eebeeaabab90de5834b32e72d2e259dc8a4a635", "parents": [ "0d6810091cdbd05efeb31654c6a41a6cbdfdd2c8", "77109cc2823f025ccd66ebd9b88fbab90437b2d8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue Oct 23 09:56:11 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue Oct 23 09:56:11 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 mlx4_core: Increase command timeout for INIT_HCA to 10 seconds\n IPoIB/cm: Use common CQ for CM send completions\n IB/uverbs: Fix checking of userspace object ownership\n IB/mlx4: Sanity check userspace send queue sizes\n IPoIB: Rewrite \"if (!likely(...))\" as \"if (unlikely(!(...)))\"\n IB/ehca: Enable large page MRs by default\n IB/ehca: Change meaning of hca_cap_mr_pgsize\n IB/ehca: Fix ehca_encode_hwpage_size() and alloc_fmr()\n IB/ehca: Fix masking error in {,re}reg_phys_mr()\n IB/ehca: Supply QP token for SRQ base QPs\n IPoIB: Use round_jiffies() for ah_reap_task\n RDMA/cma: Fix deadlock destroying listen requests\n RDMA/cma: Add locking around QP accesses\n IB/mthca: Avoid alignment traps when writing doorbells\n mlx4_core: Kill mlx4_write64_raw()\n" }, { "commit": "45711f1af6eff1a6d010703b4862e0d2b9afd056", "tree": "3d0048f46e3df9d217d56127462ebe680348bd5a", "parents": [ "78c2f0b8c285c5305b3e67b0595200541e15eb43" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Mon Oct 22 21:19:53 2007 +0200" }, "committer": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Mon Oct 22 21:19:53 2007 +0200" }, "message": "[SG] Update drivers to use sg helpers\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "77109cc2823f025ccd66ebd9b88fbab90437b2d8", "tree": "d2e4e30e8d8aa850d2fc8d1cdb1c2ee11188329a", "parents": [ "1b524963fd2d7fb20ea68df497151aa9d17fbca4" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Sun Oct 21 12:03:01 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sun Oct 21 15:06:04 2007 -0700" }, "message": "mlx4_core: Increase command timeout for INIT_HCA to 10 seconds\n\nThe current INIT_HCA firmware command timeout is sufficient for the\ndefault number of resources (QPs, CQs, etc) being allocated, but if\nthe HCA profile is modified to increase the amount of resources, then\na spurious timeout is detected and HCA initialization fails.\n\nIncrease the timeout for the INIT_HCA command to 10 seconds, which\nalso brings it into line with all the other command timeouts.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b027cacdabc3ffe2d2b97a55375203add64a5d29", "tree": "37787d8ee8d5192213bff8b8916499591e25058b", "parents": [ "e86908614f2c7fec401827e5cefd7a6ea9407f85" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sat Oct 13 14:10:50 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sat Oct 13 14:10:50 2007 -0700" }, "message": "mlx4_core: Fix infinite loop on device initialization\n \nCommit 3d73c288 (\"mlx4_core: Fix section mismatches\") introduced a\nstupid bug in device init: when some of mlx4_init_one() was split off\ninto __mlx4_init_one(), the call from the main mlx4_init_one()\nfunction was back to mlx4_init_one() rather than to __mlx4_init_one(),\nwhich leads to an obvious infinite loop if the function is every\ncalled.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "3d73c2884f45f9a297cbc956cea101405a9703f2", "tree": "2811672910ba925916e568f5275e0be630bec2ca", "parents": [ "335a64a5a958002bc238c90de695e120c3c8c120" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Oct 10 15:43:54 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Wed Oct 10 15:43:54 2007 -0700" }, "message": "mlx4_core: Fix section mismatches\n \nCommit ee49bd93 (\"mlx4_core: Reset device when internal error is\ndetected\") introduced some section mismatch problems when\nCONFIG_HOTPLUG\u003dn, because the error recovery code tears down and\nreinitializes the device after everything is loaded, which ends up\ncalling into lots of code marked __devinit and __devexit from regular\n.text. Fix this by getting rid of these now-incorrect section\nmarkers.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "2e61c646edfa013203e3428762f8d6a72e10bdea", "tree": "220800ad1947c7b4355240e45394b553963081ac", "parents": [ "76d7cc0345a037e8eea426f8abc710abd22946dd" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:18 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:18 2007 -0700" }, "message": "mlx4_core: Use mmiowb() to avoid firmware commands getting jumbled up\n\nFirmware commands are sent to the HCA by writing multiple words to a\ncommand register block. Access to this block of registers is\nserialized with a mutex. However, on large SGI systems writes to the\nregister block may be reordered within the system interconnect and\nreach the HCA in a different order than they were issued (even with\nthe mutex). Fix this by adding an mmiowb() before dropping the mutex.\n\nThis bug was observed with real workloads with the similar FW command\ncode in the mthca driver, and adding the mmiowb() as in commit\n66547550 (\"IB/mthca: Use mmiowb() to avoid firmware commands getting\njumbled up\") was confirmed to fix the problems, so we should add the\nsame fix to mlx4.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "e57ac0c297b9e1074dcb6b6ee5cd1d6407a7df14", "tree": "e2f16cec73f527e628e687709bba92c082342893", "parents": [ "8ad11fb6b0739e704953e2b0aed453bf7d75d4f6" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Tue Oct 02 09:40:13 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "message": "mlx4_core: Increase max number of QPs per multicast group to 56\n\nIncrease the number of QPs allowed per multicast group from 8 to 56.\nThis allows for one QP per core on 16-core systems, which are now\nquite common, and allows some space for future growth.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "8ad11fb6b0739e704953e2b0aed453bf7d75d4f6", "tree": "671838bc9293987efba214f2d645d7e40befc5a8", "parents": [ "d7bb58fb1c0e7264a7261c7d0304121ef9402e94" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Wed Aug 01 12:29:05 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "message": "IB/mlx4: Implement FMRs\n\nImplement FMRs for mlx4. This is an adaptation of code from mthca.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@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" }, { "commit": "d7bb58fb1c0e7264a7261c7d0304121ef9402e94", "tree": "aefbae7ee4b4b84022a7646372ddfcdb0a99e72f", "parents": [ "121964ec38d3e17b5ea6183b3c0571df2f4b0eb6" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Wed Aug 01 12:28:53 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "message": "mlx4_core: Write MTTs from CPU instead with of WRITE_MTT FW command\n\nWrite MTT entries directly to ICM from the driver (eliminating use of\nWRITE_MTT command). This reduces the number of FW commands needed to\nregister an MR by at least a factor of 2 and speeds up memory\nregistration significantly. This code will also be used to implement\nFMRs.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@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" }, { "commit": "121964ec38d3e17b5ea6183b3c0571df2f4b0eb6", "tree": "4146d90919dd8ccc2d363b879f8cdbed618a0abf", "parents": [ "cf78237d7b0e683d172e40b1e85b26ca49a3cdba" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "message": "mlx4_core: Fix meaning of dev-\u003ecaps.reserved_mtts\n\nEverything that uses caps.reserved_mtts expects it to be a count of MTT\nsegments, not MTT entries. So convert the value that the FW gives us to\na count of segments.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "cf78237d7b0e683d172e40b1e85b26ca49a3cdba", "tree": "37ab3ac0744ef2a699649e8949260301dc570f51", "parents": [ "5b0bf5e25efea77103b0ac7c8057cd56c778ef41" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:16 2007 -0700" }, "message": "mlx4_core: Reserve the correct number of MTT segments\n\nTaking ilog2(dev-\u003ecaps.reserved_mtts) to find out the order to pass to\nthe MTT buddy allocator will do the wrong thing if reserved_mtts is ever\nnot a power of 2. Be safe and use fls(dev-\u003ecaps.reserved_mtts - 1).\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "5b0bf5e25efea77103b0ac7c8057cd56c778ef41", "tree": "60fc637ca840c2a88563edc495d01f2ad75a5b69", "parents": [ "04d29b0ede242000b24cfc34cc78fbd164c47e1a" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Wed Aug 01 12:28:20 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:15 2007 -0700" }, "message": "mlx4_core: Support ICM tables in coherent memory\n\nEnable having ICM tables in coherent memory, and use coherent memory\nfor the dMPT table. This will allow writing MPT entries for MRs both\nvia the SW2HW_MPT command and also directly by the driver for FMR\nremapping without needing to flush or worry about cacheline boundaries.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@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" }, { "commit": "cd9281d873c91a01af0cb96ff0f75e9905e54403", "tree": "a639777f03924198a0c91c8884406e8ba8e40efd", "parents": [ "57cb61d587e990d556385d367589ff61f6c2c0f2" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Tue Sep 18 09:14:18 2007 +0200" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:14 2007 -0700" }, "message": "IB/mlx4: Display misc device information under /sys/class/infiniband/\n\ndisplay the following device information under /sys/class/infiniband/mlx4_X:\nboard_id, fw_ver, hw_rev, hca_type.\n\nThis patch makes this information available to userspace utilities\nsuch as ibstat and ibv_devinfo.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "ea98054fefa28b646c740baffe83c2962f8228f6", "tree": "6f554f522275520f39fcd10cb31ab44da5c75935", "parents": [ "ec2a1344ad348a789b1d9d9b32cccbef33161574" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:13 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:13 2007 -0700" }, "message": "mlx4_core: Change capability decoding: SRC-\u003eXRC\n\nThe SRC (\"scalable RC\") transport has been renamed to XRC (\"extended \nRC\"), to avoid having an abbreviation that is so easily confused with an \nabbreviation for \"source.\" Update the HCA capability decoding output to \nuse the new name.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "d7dc3ccbe4dd1d37950da0138079e61d5be06ca9", "tree": "6391db1b61ae23c34c351b617803d31dc9255804", "parents": [ "ca6de177acef8f2c7c3901ea583a263364ca7bbb" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:06 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:06 2007 -0700" }, "message": "IB/mlx4: Fix up SRQ limit_watermark endianness\n\nmlx4_srq_query() returns a big-endian 16-bit value through an int *,\nwhich screws up sparse checking. Fix this so that a CPU-endian value\nis returned.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "08fb105540f4649eaa25270ec3ada1d35406afcc", "tree": "5b2032fb7c687c7c0a6b6aba5e695233532b2f66", "parents": [ "8a68bbe31d51cc75d2acf1c9f5ceff91f7662ea4" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Tue Aug 07 16:08:28 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:05 2007 -0700" }, "message": "mlx4_core: Enable MSI-X by default\n\nRecover from MSI-X errors by automatically falling back on regular\ninterrupt, instead of asking the user to do this manually. This makes\nit possible to enable MSI-X by default, and will make it possible to\nget rid of the msi_x module option in the future.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "eaf559bf566f76887533c077d425adce847f06c8", "tree": "3ebe7ff1b45cd648a2aa190d437ba923263455c9", "parents": [ "36ce10d3e849c2f9efa3fd647058de398ee276eb" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:05 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:05 2007 -0700" }, "message": "mlx4_core: Don\u0027t free special QPs in QP number bitmap\n\nSpecial QPs are not allocated using the regular QP number bitmap, so\nwhen they are destroyed, their QP number should not be freed in the\nbitmap.\n\nFound by Dotan Barak of Mellanox.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "36ce10d3e849c2f9efa3fd647058de398ee276eb", "tree": "d9db9ddf322e00f6e106e5a9b92944e9be6dd978", "parents": [ "65d470b3ea52ee1402499d6fcb4632296452e5b1" ], "author": { "name": "Dotan Barak", "email": "dotanb@dev.mellanox.co.il", "time": "Tue Aug 07 11:18:52 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Oct 09 19:59:04 2007 -0700" }, "message": "mlx4_core: Use enum value GO_BIT_TIMEOUT_MSECS\n\nRename GO_BIT_TIMEOUT to GO_BIT_TIMEOUT_MSECS for clarity, and\nactually use it as the go bit timeout (instead of having the define\nbut then ignoring it and using a hard-coded 10 * HZ for the actual\ntimeout).\n\nSigned-off-by: Dotan Barak \u003cdotanb@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "947b2a8083a03e6fff448ce8928956015614855e", "tree": "a5dc653181c0029c11c2d45ac417c67a15573b21", "parents": [ "6958e827f187c9c5cd39af075567f74f02bf3dd1" ], "author": { "name": "Eli Cohen", "email": "eli@mellanox.co.il", "time": "Mon Aug 13 17:57:03 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Aug 13 08:47:44 2007 -0700" }, "message": "mlx4_core: Wait 1 second after reset before accessing device\n\nPut a 1000 msec delay after resetting the device before attempting to\ndo config cycles on it. Not waiting causes system hangs on some\nchipsets, e.g. Intel E7520, when the driver is loaded.\n\nSigned-off-by: Eli Cohen \u003celi@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "0172e2e14c3c1df10ec0fa31d9f41c1bc3a472f3", "tree": "3c9eb482826b1d3c40a428f4911bdd0202ac786d", "parents": [ "4e8e6ee380c3858151165d7455b4954782f145a0" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Jul 26 11:16:58 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sat Jul 28 08:30:45 2007 -0700" }, "message": "mlx4_core: Remove kfree() in mlx4_mr_alloc() error flow\n\nmlx4_mr_alloc() doesn\u0027t actually allocate mr (it just initializes the\npointer that the caller passes in), so it shouldn\u0027t free it if an\nerror occurs.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "0981582dbfae86ba0306406f1af329bb702752d2", "tree": "aae961cd3a5ca4b1e23278e70527058026048986", "parents": [ "c1f74958dbd19f6a837d887ed416688c063af529" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Fri Jul 20 21:19:43 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Fri Jul 20 21:19:43 2007 -0700" }, "message": "mlx4_core: Change command token on timeout\n \nThe FW command token is currently only updated on a command completion\nevent. This means that on command timeout, the same token will be\nreused for new command, which results in a mess if the timed out\ncommand *does* eventually complete.\n\nThis is the same change as the patch for mthca from Michael\nS. Tsirkin \u003cmst@dev.mellanox.co.il\u003e that was just merged. It seems\nsensible to avoid gratuitous differences in FW command processing\nbetween mthca and mlx4.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "c9f2ba5ed26a204a78bf23aa08a4f528f11a18f4", "tree": "e20c83f536133cd121ac8151c29853aa9d2be206", "parents": [ "8f076531cd20fbf83ed889024c8133d0c71a1fe4" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Tue Jul 17 13:11:43 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 17 20:50:50 2007 -0700" }, "message": "IB/mlx4: Increase max outstanding RDMA reads as target\n\nChange the maximum number of outstanding RDMA reads allowed as a\ntarget from 4 to 16 to per QP. This allows RDMA read operations to\npipeline better.\n\nPointed out by Dotan Barak and Sagi Rotem.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "ee49bd9397cd2b8fe7a1962505d81c1d0a1366fc", "tree": "064800f1ebcf6c2586f6727aa03c85e875b96289", "parents": [ "41179e2de6962b46d1d9f2b4437243ac740efdec" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Jul 12 17:50:45 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue Jul 17 18:37:42 2007 -0700" }, "message": "mlx4_core: Reset device when internal error is detected\n\nReset the device when an internal error is detected.\n\nAlso, detect errors by polling the error buffer rather than using\ninterrupts. This is more robust and doesn\u0027t depend on MSI-X. Remove\nthe old interrupt handler entirely, since we don\u0027t want to support two\nmechanisms for detecting internal errors.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\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": "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": "525f5f44c487201c91a11904516ee56dea93644e", "tree": "396282e71499eaa6f16b52c202310909f62300ed", "parents": [ "63019d9329628a5056b3a0c76aee955873dc25c3" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@dev.mellanox.co.il", "time": "Mon Jul 09 20:12:20 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 09 20:12:20 2007 -0700" }, "message": "mlx4_core: Include linux/mutex.h from mlx4.h\n\nmlx4.h uses struct mutex, so although \u003clinux/mutex.h\u003e seems to be pulled in\nindirectly by one of the headers it includes, the right thing to do is\nto 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": "287aa83dffd1b39859f49d73b0d67f57106de5f1", "tree": "8101259996254901fd15f8be6314efddd86e476a", "parents": [ "40738f3fcdb951fb8ade286dc1ea05812acc94db" ], "author": { "name": "Bill Nottingham", "email": "notting@redhat.com", "time": "Wed May 30 03:59:02 2007 -0400" }, "committer": { "name": "Jeff Garzik", "email": "jeff@garzik.org", "time": "Sun Jul 08 22:16:39 2007 -0400" }, "message": "drivers/net: fix comparisons of unsigned \u003c 0\n\nRecent gcc versions emit warnings when unsigned variables are compared \u003c 0 or \u003e\u003d 0.\n\nSigned-off-by: Bill Nottingham \u003cnotting@redhat.com\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "786f238e4f2f0d962831ee3228f2327133eb082a", "tree": "03e2a83ff7fcb191f5b45d24377a5d8aa9eaffb1", "parents": [ "0471448f4d017470995d8a2272dc8c06dbed3b77" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Mon Jul 02 17:36:18 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Mon Jul 02 20:41:35 2007 -0700" }, "message": "mlx4_core: Add new Mellanox device IDs\n\nAdd new IDs for PCIe gen2 devices.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\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": "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": "b2d9308ae43a9cfa56cc9682dc6ed63347a8d2bf", "tree": "cfffa29e48314d38b23e69e3db94751277ed26d4", "parents": [ "fe40900f408642e772739088d30636e2f3f7d0d8" ], "author": { "name": "Jack Morgenstein", "email": "jackm@dev.mellanox.co.il", "time": "Thu Jun 07 23:24:38 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:38 2007 -0700" }, "message": "mlx4_core: Don\u0027t set MTT address in dMPT entries with PA set\n\nIf a dMPT entry has the PA flag (direct physical address) set, then\nthe (unused) MTT base address field has to be set to 0.\n\nSigned-off-by: Jack Morgenstein \u003cjackm@dev.mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "fe40900f408642e772739088d30636e2f3f7d0d8", "tree": "e4fb317dcfd9c7145ab8a6a75f7e6c5a64e429d9", "parents": [ "3e1db334dca97df9034ebeec4657329a2b37a811" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:36 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 23:24:36 2007 -0700" }, "message": "mlx4_core: Check firmware command interface revision\n\nHCA firmware with incompatible changes to the FW commmand interface is\ncoming soon. Add a check of the interface revision during\ninitialization and bail out if the firmware advertises a revision that\nthe driver doesn\u0027t know about. This will avoid strange failures later\nif the driver goes on using the wrong interface revision. \n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\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": "2c5cb2355843f9958e19a4b243456be92f97c73b", "tree": "1c793ff7d6a1805a97c8e6d709a8b72a2df5d454", "parents": [ "b581401ed0cc83a4483ed39c00a14a60bacecc3a" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sat Jun 02 07:16:02 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 11:51:58 2007 -0700" }, "message": "mlx4_core: Free catastrophic error MSI-X interrupt with correct dev_id\n\nWe need to pass the same dev_id to free_irq() and request_irq(). When\nusing MSI-X, the MLX4_EQ_CATAS interrupt uses a different dev_id from\nthe other interrupts.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "b581401ed0cc83a4483ed39c00a14a60bacecc3a", "tree": "fa83991f1f429c0bf9876856d5255cf4a8be79fe", "parents": [ "09360d5408cb641abff0f32a172a3332f02e8e88" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 11:51:58 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 11:51:58 2007 -0700" }, "message": "mlx4_core: Initialize ctx_list and ctx_lock earlier\n\nWe may call mlx4_dispatch_event() before mlx4_register_device() is\ncalled for a device, because for example a catastrophic error happens\nimmediately after we enable interrupts. Therefore priv-\u003ectx_list and\npriv-\u003ectx_lock need to be initialized earlier.\n\nThis bug was actually exposed by the MSI-X bug that returned IRQ numbers \nto drivers in reverse order, so that the first FW command \ninterrupt looked to mlx4 like a catastrophic error.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "09360d5408cb641abff0f32a172a3332f02e8e88", "tree": "3f8b0671cf5bafa6b1f2924c63189aa9fe345ed3", "parents": [ "7244d545c123ff33a3a5bff7fcb96a07778c32d1" ], "author": { "name": "Eli Cohen", "email": "eli@mellanox.co.il", "time": "Wed May 30 13:14:31 2007 +0300" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Thu Jun 07 11:51:57 2007 -0700" }, "message": "mlx4_core: Fix CQ context layout\n\nThe reserved6 field should be 64 bits, not just 16 bits. Without\nthis, the structure does not match the hardware layout on 32-bit\narchitectures: the db_rec_addr field ends up at offset 52 instead of\noffset 56. The bug slipped by because the alignment of __be64 members\nends up putting it in the right place on x86-64.\n\nSigned-off-by: Eli Cohen \u003celi@mellanox.co.il\u003e\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "a2cb4a98f243d01f2c8d5799c764bb96ffa66c44", "tree": "8bf4e73ab877936da113969c941a51e6aef8d319", "parents": [ "d998ccce020e2cfcf11c6b57503532930ede2894" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 29 16:07:09 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 29 16:07:09 2007 -0700" }, "message": "IB/mlx4: Fix last allocated object tracking in bitmap allocator\n\nSet last allocated object to the object after the one just allocated\nbefore ORing in the extra top bits. Also handle the case where this\nwraps around.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "23c15c21d34a4b4b4d7b9a95ce498991c5339c77", "tree": "7294414512272d9a1d516c779542b9aff84a9f25", "parents": [ "65adfa911a3522c1e40e55afd472dd571dc2431b" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sat May 19 08:51:57 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sat May 19 08:51:57 2007 -0700" }, "message": "mlx4_core: Fix array overrun in dump_dev_cap_flags()\n\nDon\u0027t overrun fname[] array when decoding device flags.\n\nThis was spotted by the Coverity checker (CID 1642).\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "9cbe05c71251b51607d80256c3b3754957bb5ed8", "tree": "3ac73e809d29c47a1ecaacc616dcd7cdf6788025", "parents": [ "7a86edef395576f69d2513c62aea7c966043bc8d" ], "author": { "name": "Al Viro", "email": "viro@ftp.linux.org.uk", "time": "Tue May 15 20:36:30 2007 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue May 15 18:56:37 2007 -0700" }, "message": "missing includes in mlx4\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nAcked-by: Jeff Garzik \u003cjeff@garzik.org\u003e\nAcked-by: Roland Dreier \u003crolandd@cisco.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "de7860c3f3272086a4c3a1b4280b11ffae7c32be", "tree": "450f3cae68e5349e853c6bb201facfa60c636a90", "parents": [ "cfbf07f2a80b618c42a42c20d83647ea8fcceca0", "7c5b9ef8577bfa7b74ea58fc9ff2934ffce13532" ], "author": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue May 15 09:52:31 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Tue May 15 09:52:31 2007 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband\n\n* \u0027for-linus\u0027 of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:\n IPoIB/cm: Optimize stale connection detection\n IB/mthca: Set cleaned CQEs back to HW ownership when cleaning CQ\n IB/mthca: Fix posting \u003e255 recv WRs for Tavor\n RDMA/cma: Add check to validate that cm_id is bound to a device\n RDMA/cma: Fix synchronization with device removal in cma_iw_handler\n RDMA/cma: Simplify device removal handling code\n IB/ehca: Disable scaling code by default, bump version number\n IB/ehca: Beautify sysfs attribute code and fix compiler warnings\n IB/ehca: Remove _irqsave, move #ifdef\n IB/ehca: Fix AQP0/1 QP number\n IB/ehca: Correctly set GRH mask bit in ehca_modify_qp()\n IB/ehca: Serialize hypervisor calls in ehca_register_mr()\n IB/ipath: Shadow the gpio_mask register\n IB/mlx4: Fix uninitialized spinlock for 32-bit archs\n mlx4_core: Remove unused doorbell_lock\n net: Trivial MLX4_DEBUG dependency fix.\n" }, { "commit": "20eebcf09c2d329e4dcdd765634c0a524195e16d", "tree": "d4044790aca3555854d90ed005a603d0fb23f1e0", "parents": [ "f550d94cf9f86bc54e31dae2aee1a03d678c6e7f" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sun May 13 08:54:18 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Sun May 13 08:54:18 2007 -0700" }, "message": "mlx4_core: Remove unused doorbell_lock\n\nstruct mlx4_priv.doorbell_lock is never used, so delete it.\n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" }, { "commit": "4093785dd16d0b7058eb92331bfd7b40b12151f3", "tree": "d1b48f39599c4f66497485639cb86f2a72689c72", "parents": [ "1c8e039a4bce0dae0d49c5df451f338a5d19363a" ], "author": { "name": "Andrew Morton", "email": "akpm@linux-foundation.org", "time": "Thu May 10 22:53:01 2007 -0700" }, "committer": { "name": "Jeff Garzik", "email": "jeff@garzik.org", "time": "Fri May 11 17:53:31 2007 -0400" }, "message": "mlx4: don\u0027t use deprecated IRQ flags\n\nCc: Jeff Garzik \u003cjeff@garzik.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "225c7b1feef1b41170f7037a5b10a65cd8a42c54", "tree": "702a0a2cbba7f1c5b2949d236b4463d486204fdc", "parents": [ "1bf66a30421ca772820f489d88c16d0c430d6a67" ], "author": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 08 18:00:38 2007 -0700" }, "committer": { "name": "Roland Dreier", "email": "rolandd@cisco.com", "time": "Tue May 08 18:00:38 2007 -0700" }, "message": "IB/mlx4: Add a driver Mellanox ConnectX InfiniBand adapters\n\nAdd an InfiniBand driver for Mellanox ConnectX adapters. Because\nthese adapters can also be used as ethernet NICs and Fibre Channel \nHBAs, the driver is split into two modules: \n \n mlx4_core: Handles low-level things like device initialization and \n processing firmware commands. Also controls resource allocation \n so that the InfiniBand, ethernet and FC functions can share a \n device without stepping on each other. \n \n mlx4_ib: Handles InfiniBand-specific things; plugs into the \n InfiniBand midlayer. \n\nSigned-off-by: Roland Dreier \u003crolandd@cisco.com\u003e\n" } ] }