)]}' { "log": [ { "commit": "1096d63d8e7d226630706e15648705d0187787e4", "tree": "0233951b529ee7b537d527e6d57c8e5d80d48ae7", "parents": [ "ea866e6526b8a2ead92875732d41b26fdb470312" ], "author": { "name": "Robert Jennings", "email": "rcj@linux.vnet.ibm.com", "time": "Thu Jul 24 04:34:52 2008 +1000" }, "committer": { "name": "Benjamin Herrenschmidt", "email": "benh@kernel.crashing.org", "time": "Fri Jul 25 15:44:44 2008 +1000" }, "message": "ibmveth: enable driver for CMO\n\nEnable ibmveth for Cooperative Memory Overcommitment (CMO). For this driver\nit means calculating a desired amount of IO memory based on the current MTU\nand updating this value with the bus when MTU changes occur. Because DMA\nmappings can fail, we have added a bounce buffer for temporary cases where\nthe driver can not map IO memory for the buffer pool.\n\nThe following changes are made to enable the driver for CMO:\n * DMA mapping errors will not result in error messages if entitlement has\n been exceeded and resources were not available.\n * DMA mapping errors are handled gracefully, ibmveth_replenish_buffer_pool()\n is corrected to check the return from dma_map_single and fail gracefully.\n * The driver will have a get_desired_dma function defined to function\n in a CMO environment.\n * When the MTU is changed, the driver will update the device IO entitlement\n\nSigned-off-by: Robert Jennings \u003crcj@linux.vnet.ibm.com\u003e\nSigned-off-by: Brian King \u003cbrking@linux.vnet.ibm.com\u003e\nSigned-off-by: Santiago Leon \u003csantil@us.ibm.com\u003e\nAcked-by: Paul Mackerras \u003cpaulus@samba.org\u003e\nSigned-off-by: Benjamin Herrenschmidt \u003cbenh@kernel.crashing.org\u003e\n" }, { "commit": "79ef4a4dd44cd4f9942975b0f625bd01549a2aa9", "tree": "177d1e948721db288f2f13487e42d47cd9acef83", "parents": [ "3449a2ab31681420515e242920e755262b4f41e9" ], "author": { "name": "Brian King", "email": "brking@linux.vnet.ibm.com", "time": "Fri Aug 17 09:16:56 2007 -0500" }, "committer": { "name": "David S. Miller", "email": "davem@sunset.davemloft.net", "time": "Wed Oct 10 16:50:46 2007 -0700" }, "message": "ibmveth: Remove use of bitfields\n\nRemoves the use of bitfields from the ibmveth driver. This results\nin slightly smaller object code.\n\nSigned-off-by: Brian King \u003cbrking@linux.vnet.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "3449a2ab31681420515e242920e755262b4f41e9", "tree": "71ab1800bd82b3337b2a058d45b0e4b328aec4e0", "parents": [ "ddbb4de9672097da2c0f19c6ebca0ebb5672e9b8" ], "author": { "name": "Brian King", "email": "brking@linux.vnet.ibm.com", "time": "Fri Aug 17 09:16:49 2007 -0500" }, "committer": { "name": "David S. Miller", "email": "davem@sunset.davemloft.net", "time": "Wed Oct 10 16:50:45 2007 -0700" }, "message": "ibmveth: Remove dead frag processing code\n\nRemoves dead frag processing code from ibmveth. Since NETIF_F_SG was\nnot set, this code was never executed. Also, since the ibmveth\ninterface can only handle 6 fragments, core networking code would need\nto be modified in order to efficiently enable this support.\n\nSigned-off-by: Brian King \u003cbrking@linux.vnet.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "5fc7e01cb77132f96e171a37f9f792270b1603f6", "tree": "44c4e1926a803ef1e53f5a1985f976ba6dd35171", "parents": [ "f4ff28720f45354573dcf4e0eb5a2dc5452cb3e1" ], "author": { "name": "Brian King", "email": "brking@linux.vnet.ibm.com", "time": "Fri Aug 17 09:16:31 2007 -0500" }, "committer": { "name": "David S. Miller", "email": "davem@sunset.davemloft.net", "time": "Wed Oct 10 16:50:43 2007 -0700" }, "message": "ibmveth: Implement ethtool hooks to enable/disable checksum offload\n\nThis patch adds the appropriate ethtool hooks to allow for enabling/disabling\nof hypervisor assisted checksum offload for TCP.\n\nSigned-off-by: Brian King \u003cbrking@linux.vnet.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "f4ff28720f45354573dcf4e0eb5a2dc5452cb3e1", "tree": "bdd8f33e738962c44f1172a3fd3a3aebb0246f61", "parents": [ "df950828b0ee51ff63c49c67d561bfd3d6096788" ], "author": { "name": "Brian King", "email": "brking@linux.vnet.ibm.com", "time": "Sat Sep 15 13:36:07 2007 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@sunset.davemloft.net", "time": "Wed Oct 10 16:50:43 2007 -0700" }, "message": "ibmveth: Enable TCP checksum offload\n\nThis patchset enables TCP checksum offload support for IPV4\non ibmveth. This completely eliminates the generation and checking of\nthe checksum for packets that are completely virtual and never\ntouch a physical network. A simple TCP_STREAM netperf run on\na virtual network with maximum mtu set yielded a ~30% increase\nin throughput. This feature is enabled by default on systems that\nsupport it, but can be disabled with a module option.\n\nSigned-off-by: Brian King \u003cbrking@linux.vnet.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "bea3348eef27e6044b6161fd04c3152215f96411", "tree": "f0990b263e5ce42505d290a4c346fe990bcd4c33", "parents": [ "dde4e47e8fe333a5649a3fa0e7db1fa7c08d6158" ], "author": { "name": "Stephen Hemminger", "email": "shemminger@linux-foundation.org", "time": "Wed Oct 03 16:41:36 2007 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@sunset.davemloft.net", "time": "Wed Oct 10 16:47:45 2007 -0700" }, "message": "[NET]: Make NAPI polling independent of struct net_device objects.\n\nSeveral devices have multiple independant RX queues per net\ndevice, and some have a single interrupt doorbell for several\nqueues.\n\nIn either case, it\u0027s easier to support layouts like that if the\nstructure representing the poll is independant from the net\ndevice itself.\n\nThe signature of the -\u003epoll() call back goes from:\n\n\tint foo_poll(struct net_device *dev, int *budget)\n\nto\n\n\tint foo_poll(struct napi_struct *napi, int budget)\n\nThe caller is returned the number of RX packets processed (or\nthe number of \"NAPI credits\" consumed if you want to get\nabstract). The callee no longer messes around bumping\ndev-\u003equota, *budget, etc. because that is all handled in the\ncaller upon return.\n\nThe napi_struct is to be embedded in the device driver private data\nstructures.\n\nFurthermore, it is the driver\u0027s responsibility to disable all NAPI\ninstances in it\u0027s -\u003estop() device close handler. Since the\nnapi_struct is privatized into the driver\u0027s private data structures,\nonly the driver knows how to get at all of the napi_struct instances\nit may have per-device.\n\nWith lots of help and suggestions from Rusty Russell, Roland Dreier,\nMichael Chan, Jeff Garzik, and Jamal Hadi Salim.\n\nBug fixes from Thomas Graf, Roland Dreier, Peter Zijlstra,\nJoseph Fannin, Scott Wood, Hans J. Koch, and Michael Chan.\n\n[ Ported to current tree and all drivers converted. Integrated\n Stephen\u0027s follow-on kerneldoc additions, and restored poll_list\n handling to the old style to fix mutual exclusion issues. -DaveM ]\n\nSigned-off-by: Stephen Hemminger \u003cshemminger@linux-foundation.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "76b9cfccb39f542ce48a79b9ad50af25e422506c", "tree": "25424ff8bdca2d079eae12ff169ef48d39354e9e", "parents": [ "c21723edd5f6c288ab613ed658b7140e07fc7209" ], "author": { "name": "Brian King", "email": "brking@linux.vnet.ibm.com", "time": "Fri Aug 03 13:55:19 2007 +1000" }, "committer": { "name": "Jeff Garzik", "email": "jeff@garzik.org", "time": "Tue Aug 07 17:34:13 2007 -0400" }, "message": "ibmveth: Fix rx pool deactivate oops\n\nThis fixes the following oops which can occur when trying to deallocate\nreceive buffer pools using sysfs with the ibmveth driver.\n\nNIP: d00000000024f954 LR: d00000000024fa58 CTR: c0000000000d7478\nREGS: c00000000ffef9f0 TRAP: 0300 Not tainted (2.6.22-ppc64)\nMSR: 8000000000009032 \u003cEE,ME,IR,DR\u003e CR: 24242442 XER: 00000010\nDAR: 00000000000007f0, DSISR: 0000000042000000\nTASK \u003d c000000002f91360[2967] \u0027bash\u0027 THREAD: c00000001398c000 CPU: 2\nGPR00: 0000000000000000 c00000000ffefc70 d000000000262d30 c00000001c4087a0\nGPR04: 00000003000000fe 0000000000000000 000000000000000f c000000000579d80\nGPR08: 0000000000365688 c00000001c408998 00000000000007f0 0000000000000000\nGPR12: d000000000251e88 c000000000579d80 00000000200957ec 0000000000000000\nGPR16: 00000000100b8808 00000000100feb30 0000000000000000 0000000010084828\nGPR20: 0000000000000000 000000001014d4d0 0000000000000010 c00000000ffefeb0\nGPR24: c00000001c408000 0000000000000000 c00000001c408000 00000000ffffb054\nGPR28: 00000000000000fe 0000000000000003 d000000000262700 c00000001c4087a0\nNIP [d00000000024f954] .ibmveth_remove_buffer_from_pool+0x38/0x108 [ibmveth]\nLR [d00000000024fa58] .ibmveth_rxq_harvest_buffer+0x34/0x78 [ibmveth]\nCall Trace:\n[c00000000ffefc70] [c0000000000280a8] .dma_iommu_unmap_single+0x14/0x28 (unreliable)\n[c00000000ffefd00] [d00000000024fa58] .ibmveth_rxq_harvest_buffer+0x34/0x78 [ibmveth]\n[c00000000ffefd80] [d000000000250e40] .ibmveth_poll+0xd8/0x434 [ibmveth]\n[c00000000ffefe40] [c00000000032da8c] .net_rx_action+0xdc/0x248\n[c00000000ffefef0] [c000000000068b4c] .__do_softirq+0xa8/0x164\n[c00000000ffeff90] [c00000000002789c] .call_do_softirq+0x14/0x24\n[c00000001398f6f0] [c00000000000c04c] .do_softirq+0x68/0xac\n[c00000001398f780] [c000000000068ca0] .irq_exit+0x54/0x6c\n[c00000001398f800] [c00000000000c8e4] .do_IRQ+0x170/0x1ac\n[c00000001398f890] [c000000000004790] hardware_interrupt_entry+0x18/0x1c\n Exception: 501 at .plpar_hcall_norets+0x24/0x94\n LR \u003d .veth_pool_store+0x15c/0x298 [ibmveth]\n[c00000001398fb80] [d000000000250b2c] .veth_pool_store+0x5c/0x298 [ibmveth] (unreliable)\n[c00000001398fc30] [c000000000145530] .sysfs_write_file+0x140/0x1d8\n[c00000001398fcf0] [c0000000000de89c] .vfs_write+0x120/0x208\n[c00000001398fd90] [c0000000000df2c8] .sys_write+0x4c/0x8c\n[c00000001398fe30] [c0000000000086ac] syscall_exit+0x0/0x40\nInstruction dump:\nfba1ffe8 fbe1fff8 789d0022 f8010010 f821ff71 789c0020 1d3d00a8 7b8a1f24\n38000000 7c7f1b78 7d291a14 e9690128 \u003c7c0a592a\u003e e8030000 e9690120 80a90100\n\nSigned-off-by: Brian King \u003cbrking@linux.vnet.ibm.com\u003e\nSigned-off-by: Paul Mackerras \u003cpaulus@samba.org\u003e\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "7c719871ff4d5f15b71f0138d08b758281b58631", "tree": "a907bb8ac3b84264a80b5ce9525369dd69684271", "parents": [ "96289b07eb319ab3f64db3f0d981970aa1d60a60" ], "author": { "name": "Benjamin Herrenschmidt", "email": "benh@kernel.crashing.org", "time": "Sat Nov 11 17:24:42 2006 +1100" }, "committer": { "name": "Paul Mackerras", "email": "paulus@samba.org", "time": "Mon Dec 04 20:38:06 2006 +1100" }, "message": "[PATCH] ibmveth: Remove ibmveth \"liobn\" field\n\nRemove the now unused \"liobn\" field in ibmveth which also avoids\nhaving insider knowledge of the iommu table in that driver.\n\nSigned-off-by: Benjamin Herrenschmidt \u003cbenh@kernel.crashing.org\u003e\nSigned-off-by: Santiago Leon \u003csantil@us.ibm.com\u003e\nSigned-off-by: Paul Mackerras \u003cpaulus@samba.org\u003e\n" }, { "commit": "b9377ffc3a03cde558d76349a262a1adbb6d3112", "tree": "c61fcdb732d06c64b9c5634953e46cefdf6af846", "parents": [ "57cad8084e0837e0f2c97da789ec9b3f36809be9" ], "author": { "name": "Anton Blanchard", "email": "anton@samba.org", "time": "Wed Jul 19 08:01:28 2006 +1000" }, "committer": { "name": "Paul Mackerras", "email": "paulus@samba.org", "time": "Tue Aug 01 16:19:15 2006 +1000" }, "message": "[POWERPC] clean up pseries hcall interfaces\n\nOur pseries hcall interfaces are out of control:\n\n\tplpar_hcall_norets\n\tplpar_hcall\n\tplpar_hcall_8arg_2ret\n\tplpar_hcall_4out\n\tplpar_hcall_7arg_7ret\n\tplpar_hcall_9arg_9ret\n\nCreate 3 interfaces to cover all cases:\n\n\tplpar_hcall_norets:\t7 arguments no returns\n\tplpar_hcall:\t\t6 arguments 4 returns\n\tplpar_hcall9:\t\t9 arguments 9 returns\n\nThere are only 2 cases in the kernel that need plpar_hcall9, hopefully\nwe can keep it that way.\n\nPass in a buffer to stash return parameters so we avoid the \u0026dummy1,\n\u0026dummy2 madness.\n\nSigned-off-by: Anton Blanchard \u003canton@samba.org\u003e\n--\nSigned-off-by: Paul Mackerras \u003cpaulus@samba.org\u003e\n" }, { "commit": "cb18bd40030c879cd93fef02fd579f74dbab473d", "tree": "fdc2d097988ab89fe69d28e36e83b0bad802ad9b", "parents": [ "5d33eebee83784f5f03bc3861fa92ee5cd831922" ], "author": { "name": "Mike Kravetz", "email": "kravetz@us.ibm.com", "time": "Thu Jul 20 23:39:51 2006 -0700" }, "committer": { "name": "Paul Mackerras", "email": "paulus@samba.org", "time": "Mon Jul 31 15:56:35 2006 +1000" }, "message": "[POWERPC] Instrument Hypervisor Calls: merge headers\n\nMove all the Hypervisor call definitions to to a single header file.\n\nSigned-off-by: Mike Kravetz \u003ckravetz@us.ibm.com\u003e\nSigned-off-by: Paul Mackerras \u003cpaulus@samba.org\u003e\n" }, { "commit": "d7fbeba60b4021dfaf5d315884fbf7672b4ae87b", "tree": "1da0475fa710af806a2a83d6ee9f6d16c144a028", "parents": [ "860f242eb5340d0b0cfe243cb86b2a98f92e8b91" ], "author": { "name": "Jeff Garzik", "email": "jeff@garzik.org", "time": "Wed May 24 01:31:14 2006 -0400" }, "committer": { "name": "Jeff Garzik", "email": "jeff@garzik.org", "time": "Wed May 24 01:31:14 2006 -0400" }, "message": "[netdrvr ibmlana, ibmveth] trim trailing whitespace\n" }, { "commit": "860f242eb5340d0b0cfe243cb86b2a98f92e8b91", "tree": "286d64b4acfc392bcb926a6f5f7bfb311b0d3efc", "parents": [ "7b32a312895c00ff03178e49db8b651ee1e48178" ], "author": { "name": "Santiago Leon", "email": "santil@us.ibm.com", "time": "Tue Apr 25 11:19:59 2006 -0500" }, "committer": { "name": "Jeff Garzik", "email": "jeff@garzik.org", "time": "Wed May 24 01:30:37 2006 -0400" }, "message": "[PATCH] ibmveth change buffer pools dynamically\n\nThis patch provides a sysfs interface to change some properties of the\nibmveth buffer pools (size of the buffers, number of buffers per pool,\nand whether a pool is active). Ethernet drivers use ethtool to provide\nthis type of functionality. However, the buffers in the ibmveth driver\ncan have an arbitrary size (not only regular, mini, and jumbo which are\nthe only sizes that ethtool can change), and also ibmveth can have an\narbitrary number of buffer pools\n\nUnder heavy load we have seen dropped packets which obviously kills TCP\nperformance. We have created several fixes that mitigate this issue,\nbut we definitely need a way of changing the number of buffers for an\nadapter dynamically. Also, changing the size of the buffers allows\nusers to change the MTU to something big (bigger than a jumbo frame)\ngreatly improving performance on partition to partition transfers.\n\nThe patch creates directories pool1...pool4 in the device directory in\nsysfs, each with files: num, size, and active (which default to the\nvalues in the mainline version).\n\nComments and suggestions are welcome...\n--\nSantiago A. Leon\nPower Linux Development\nIBM Linux Technology Center\n\nSigned-off-by: Jeff Garzik \u003cjeff@garzik.org\u003e\n" }, { "commit": "60296d9e4be1cd9e096f7804ce6e839e0cbd97cf", "tree": "40372d93b1c99a0e488adedbbd8c843c26807058", "parents": [ "e2adbcb480992de8a01acf9218e8bbd9b507fc6f" ], "author": { "name": "Santiago Leon", "email": "santil@us.ibm.com", "time": "Wed Oct 26 10:47:16 2005 -0600" }, "committer": { "name": "Jeff Garzik", "email": "jgarzik@pobox.com", "time": "Fri Oct 28 16:07:30 2005 -0400" }, "message": "[PATCH] ibmveth lockless TX\n\nThis patch adds the lockless TX feature to the ibmveth driver. The\nhypervisor has its own locking so the only change that is necessary is\nto protect the statistics counters.\n\nSigned-off-by: Santiago Leon \u003csantil@us.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjgarzik@pobox.com\u003e\n" }, { "commit": "e2adbcb480992de8a01acf9218e8bbd9b507fc6f", "tree": "92465e753d0221c0c54862994735a7fe078200fa", "parents": [ "b6d35182fe62e57d368062adcc880ca35119d88e" ], "author": { "name": "Santiago Leon", "email": "santil@us.ibm.com", "time": "Wed Oct 26 10:47:08 2005 -0600" }, "committer": { "name": "Jeff Garzik", "email": "jgarzik@pobox.com", "time": "Fri Oct 28 16:07:30 2005 -0400" }, "message": "[PATCH] ibmveth fix buffer replenishing\n\nThis patch removes the allocation of RX skb\u0027s buffers from a workqueue\nto be called directly at RX processing time. This change was suggested\nby Dave Miller when the driver was starving the RX buffers and\ndeadlocking under heavy traffic:\n\n\u003e Allocating RX SKBs via tasklet is, IMHO, the worst way to\n\u003e do it. It is no surprise that there are starvation cases.\n\u003e\n\u003e If tasklets or work queues get delayed in any way, you lose,\n\u003e and it\u0027s very easy for a card to catch up with the driver RX\u0027ing\n\u003e packets very fast, no matter how aggressive you make the\n\u003e replenishing. By the time you detect that you need to be\n\u003e \"more aggressive\" it is already too late.\n\u003e The only pseudo-reliable way is to allocate at RX processing time.\n\u003e\n\nSigned-off-by: Santiago Leon \u003csantil@us.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjgarzik@pobox.com\u003e\n" }, { "commit": "b6d35182fe62e57d368062adcc880ca35119d88e", "tree": "dd7767a40490d2d532cda4d35a18f8b8e614ab19", "parents": [ "0abe791e94033b727f2b55670c2966f3d3d3cf70" ], "author": { "name": "Santiago Leon", "email": "santil@us.ibm.com", "time": "Wed Oct 26 10:47:01 2005 -0600" }, "committer": { "name": "Jeff Garzik", "email": "jgarzik@pobox.com", "time": "Fri Oct 28 16:07:30 2005 -0400" }, "message": "[PATCH] ibmveth fix buffer pool management\n\nThis patch changes the way the ibmveth driver handles the receive\nbuffers. The old code mallocs and maps all the buffers in the pools\nregardless of MTU size and it also limits the number of buffer pools to\nthree. This patch makes the driver malloc and map the buffers necessary\nto support the current MTU. It also changes the hardcoded names of the\nbuffer pool number, size, and elements to arrays to make it easier to\nchange (with the hope of making them runtime parameters in the future).\n\nSigned-off-by: Santiago Leon \u003csantil@us.ibm.com\u003e\nSigned-off-by: Jeff Garzik \u003cjgarzik@pobox.com\u003e\n" }, { "commit": "1da177e4c3f41524e886b7f1b8a0c1fc7321cac2", "tree": "0bba044c4ce775e45a88a51686b5d9f90697ea9d", "parents": [], "author": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "message": "Linux-2.6.12-rc2\n\nInitial git repository build. I\u0027m not bothering with the full history,\neven though we have it. We can create a separate \"historical\" git\narchive of that later if we want to, and in the meantime it\u0027s about\n3.2GB when imported into git - space that would just make the early\ngit days unnecessarily complicated, when we don\u0027t have a lot of good\ninfrastructure for it.\n\nLet it rip!\n" } ] }