)]}' { "log": [ { "commit": "2d61ba95034f1abbdec7729d52c740870a5eddb6", "tree": "ec51d4cba4240af6011f4fabfda06a21532761da", "parents": [ "0b22bd0ba0d500ab40e7673c591c594ec5447342" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Sun Oct 25 15:28:53 2009 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Oct 29 08:50:37 2009 +1030" }, "message": "virtio: order used ring after used index read\n\nOn SMP guests, reads from the ring might bypass used index reads. This\ncauses guest crashes because host writes to used index to signal ring\ndata readiness. Fix this by inserting rmb before used ring reads.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: stable@kernel.org\n" }, { "commit": "0b22bd0ba0d500ab40e7673c591c594ec5447342", "tree": "09146fc159cc456d79f20f5fb66cbc7bfa8c9871", "parents": [ "964fe080d94db82a3268443e9b9ece4c60246414" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Thu Oct 22 15:06:06 2009 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Oct 29 08:50:36 2009 +1030" }, "message": "virtio-pci: fix per-vq MSI-X request logic\n\nCommit f68d24082e22ccee3077d11aeb6dc5354f0ca7f1\nin 2.6.32-rc1 broke requesting IRQs for per-VQ MSI-X vectors:\n- vector number was used instead of the vector itself\n- we try to request an IRQ for VQ which does not\n have a callback handler\n\nThis is a regression that causes warnings in kernel log,\npotentially lower performance as we need to scan vq list,\nand might cause system failure if the interrupt\nrequested is in fact needed by another system.\n\nThis was not noticed earlier because in most cases\nwe were falling back on shared interrupt for all vqs.\n\nThe warnings often look like this:\n\nvirtio-pci 0000:00:03.0: irq 26 for MSI/MSI-X\nvirtio-pci 0000:00:03.0: irq 27 for MSI/MSI-X\nvirtio-pci 0000:00:03.0: irq 28 for MSI/MSI-X\nIRQ handler type mismatch for IRQ 1\ncurrent handler: i8042\nPid: 2400, comm: modprobe Tainted: G W\n2.6.32-rc3-11952-gf3ed8d8-dirty #1\nCall Trace:\n [\u003cffffffff81072aed\u003e] ? __setup_irq+0x299/0x304\n [\u003cffffffff81072ff3\u003e] ? request_threaded_irq+0x144/0x1c1\n [\u003cffffffff813455af\u003e] ? vring_interrupt+0x0/0x30\n [\u003cffffffff81346598\u003e] ? vp_try_to_find_vqs+0x583/0x5c7\n [\u003cffffffffa0015188\u003e] ? skb_recv_done+0x0/0x34 [virtio_net]\n [\u003cffffffff81346609\u003e] ? vp_find_vqs+0x2d/0x83\n [\u003cffffffff81345d00\u003e] ? vp_get+0x3c/0x4e\n [\u003cffffffffa0016373\u003e] ? virtnet_probe+0x2f1/0x428 [virtio_net]\n [\u003cffffffffa0015188\u003e] ? skb_recv_done+0x0/0x34 [virtio_net]\n [\u003cffffffffa00150d8\u003e] ? skb_xmit_done+0x0/0x39 [virtio_net]\n [\u003cffffffff8110ab92\u003e] ? sysfs_do_create_link+0xcb/0x116\n [\u003cffffffff81345cc2\u003e] ? vp_get_status+0x14/0x16\n [\u003cffffffff81345464\u003e] ? virtio_dev_probe+0xa9/0xc8\n [\u003cffffffff8122b11c\u003e] ? driver_probe_device+0x8d/0x128\n [\u003cffffffff8122b206\u003e] ? __driver_attach+0x4f/0x6f\n [\u003cffffffff8122b1b7\u003e] ? __driver_attach+0x0/0x6f\n [\u003cffffffff8122a9f9\u003e] ? bus_for_each_dev+0x43/0x74\n [\u003cffffffff8122a374\u003e] ? bus_add_driver+0xea/0x22d\n [\u003cffffffff8122b4a3\u003e] ? driver_register+0xa7/0x111\n [\u003cffffffffa001a000\u003e] ? init+0x0/0xc [virtio_net]\n [\u003cffffffff81009051\u003e] ? do_one_initcall+0x50/0x148\n [\u003cffffffff8106e117\u003e] ? sys_init_module+0xc5/0x21a\n [\u003cffffffff8100af02\u003e] ? system_call_fastpath+0x16/0x1b\nvirtio-pci 0000:00:03.0: irq 26 for MSI/MSI-X\nvirtio-pci 0000:00:03.0: irq 27 for MSI/MSI-X\n\nReported-by: Marcelo Tosatti \u003cmtosatti@redhat.com\u003e\nReported-by: Shirley Ma \u003cxma@us.ibm.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "1e65175c2c73742495f0e5ca52658539a65825db", "tree": "4e6538fa79a04ff5f6666cde1e8b84f7857192a8", "parents": [ "3225beaba05d4f06087593f5e903ce867b6e118a" ], "author": { "name": "Uwe Kleine-König", "email": "u.kleine-koenig@pengutronix.de", "time": "Thu Oct 01 10:28:33 2009 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Oct 22 16:39:31 2009 +1030" }, "message": "move virtballoon_remove to .devexit.text\n\nThe function virtballoon_remove is used only wrapped by __devexit_p so\ndefine it using __devexit.\n\nSigned-off-by: Uwe Kleine-König \u003cu.kleine-koenig@pengutronix.de\u003e\nAcked-by: Sam Ravnborg \u003csam@ravnborg.org\u003e\nAcked-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "e95646c3ec33c8ec0693992da4332a6b32eb7e31", "tree": "7855767cc4dccdefc4ea64584cc01b64f92176f3", "parents": [ "f8b12e513b953aebf30f8ff7d2de9be7e024dbbe" ], "author": { "name": "Christian Borntraeger", "email": "borntraeger@de.ibm.com", "time": "Wed Sep 30 11:17:21 2009 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Oct 22 16:39:28 2009 +1030" }, "message": "virtio: let header files include virtio_ids.h\n\nRusty,\n\ncommit 3ca4f5ca73057a617f9444a91022d7127041970a\n virtio: add virtio IDs file\nmoved all device IDs into a single file. While the change itself is\na very good one, it can break userspace applications. For example\nif a userspace tool wanted to get the ID of virtio_net it used to\ninclude virtio_net.h. This does no longer work, since virtio_net.h\ndoes not include virtio_ids.h.\nThis patch moves all \"#include \u003clinux/virtio_ids.h\u003e\" from the C\nfiles into the header files, making the header files compatible with\nthe old ones.\n\nIn addition, this patch exports virtio_ids.h to userspace.\n\nCC: Fernando Luis Vazquez Cao \u003cfernando@oss.ntt.co.jp\u003e\nSigned-off-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "3ca4f5ca73057a617f9444a91022d7127041970a", "tree": "382659da29ad0a2e9878df6c7c4204913ab642e7", "parents": [ "3a20210dc26bbfff3bbb48bb22d2846240b71d8f" ], "author": { "name": "Fernando Luis Vazquez Cao", "email": "fernando@oss.ntt.co.jp", "time": "Fri Jul 31 15:25:56 2009 +0900" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Wed Sep 23 22:26:32 2009 +0930" }, "message": "virtio: add virtio IDs file\n\nVirtio IDs are spread all over the tree which makes assigning new IDs\nbothersome. Putting them together should make the process less error-prone.\n\nSigned-off-by: Fernando Luis Vazquez Cao \u003cfernando@oss.ntt.co.jp\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "3c1b27d5043086a485f8526353ae9fe37bfa1065", "tree": "e3b6eda3c66bcd1fc3af6e7fa6e4e3af77459474", "parents": [ "f68d24082e22ccee3077d11aeb6dc5354f0ca7f1" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Wed Sep 23 22:26:31 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Wed Sep 23 22:26:31 2009 +0930" }, "message": "virtio: make add_buf return capacity remaining\n\nThis API change means that virtio_net can tell how much capacity\nremains for buffers. It\u0027s necessarily fuzzy, since\nVIRTIO_RING_F_INDIRECT_DESC means we can fit any number of descriptors\nin one, *if* we can kmalloc.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Dinesh Subhraveti \u003cdineshs@us.ibm.com\u003e\n" }, { "commit": "f68d24082e22ccee3077d11aeb6dc5354f0ca7f1", "tree": "22e2ab313dda8217926b0ddf0c760554391ee092", "parents": [ "ebc79c4f8da0f92efa968e0328f32334a2ce80cf" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Wed Sep 23 22:26:29 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Wed Sep 23 22:26:31 2009 +0930" }, "message": "virtio_pci: minor MSI-X cleanups\n\n1) Rename vp_request_vectors to vp_request_msix_vectors, and take\n non-MSI-X case out to caller.\n2) Comment weird pci_enable_msix API\n3) Rename vp_find_vq to setup_vq.\n4) Fix spaces to tabs\n5) Make nvectors calc internal to vp_try_to_find_vqs()\n6) Rename vector to msix_vector for more clarity.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: \"Michael S. Tsirkin\" \u003cmst@redhat.com\u003e\n" }, { "commit": "e969fed542cae08cb11d666efac4f7c5d624d09f", "tree": "7de7cb45ba6b3a28ef0054df71c7dc7676d4af65", "parents": [ "f6c82507030d61e15928d5cad946d3eac1c4a384" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Sun Jul 26 15:48:08 2009 +0300" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Jul 30 16:03:45 2009 +0930" }, "message": "virtio: refactor find_vqs\n\nThis refactors find_vqs, making it more readable and robust, and fixing\ntwo regressions from 2.6.30:\n- double free_irq causing BUG_ON on device removal\n- probe failure when vq can\u0027t be assigned to msi-x vector\n (reported on old host kernels)\n\nTested-by: Amit Shah \u003camit.shah@redhat.com\u003e\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "f6c82507030d61e15928d5cad946d3eac1c4a384", "tree": "5c0dd65a04424dbcf985a4104a8a244085e2f179", "parents": [ "ff52c3fc7188855ede75d87b022271f0da309e5b" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Sun Jul 26 15:48:01 2009 +0300" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Jul 30 16:03:44 2009 +0930" }, "message": "virtio: delete vq from list\n\nThis makes delete vq the reverse of find vq.\nThis is required to make it possible to retry find_vqs\nafter a failure, otherwise the list gets corrupted.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "ff52c3fc7188855ede75d87b022271f0da309e5b", "tree": "e10333755598cfbee6676f75dd87168ecdf2d5ce", "parents": [ "8ef562d112c82ec539775698f8b63ac5ec1bd766" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Thu Jul 23 14:57:37 2009 +0300" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Thu Jul 30 16:03:44 2009 +0930" }, "message": "virtio: fix memory leak on device removal\n\nMake vp_free_vectors do the reverse of vq_request_vectors.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "4b892e6582e3a4fe01f623aea386907270d5bf83", "tree": "82214468d537c8cf90b1fc0195bf33c688db3bfd", "parents": [ "d9ecdea7ed7467db32ec160f4eca46c279255606" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Tue Jul 07 08:26:45 2009 +0100" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 17 21:47:47 2009 +0930" }, "message": "virtio-pci: correctly unregister root device on error\n\nIf pci_register_driver() fails we\u0027re incorrectly unregistering the root\ndevice with device_unregister() rather than root_device_unregister().\n\nReported-by: Don Zickus \u003cdzickus@redhat.com\u003e\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "e3353853730eb99c56b7b0aed1667d51c0e3699a", "tree": "d0da26c34a8ca12a57662830fd698ca6696f3805", "parents": [ "c89e80168ba1ed37627fe03116b0cf8474dcb7e0" ], "author": { "name": "Christian Borntraeger", "email": "borntraeger@de.ibm.com", "time": "Tue May 26 15:46:10 2009 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:40 2009 +0930" }, "message": "virtio: enhance id_matching for virtio drivers\n\nThis patch allows a virtio driver to use VIRTIO_DEV_ANY_ID for the\ndevice id. This will be used by a test module that can be bound to\nany virtio device.\n\nSigned-off-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "c89e80168ba1ed37627fe03116b0cf8474dcb7e0", "tree": "3314d284cca77dc81913877d1cf58838186e7e85", "parents": [ "594de1dd6449f79c99e1ba4577ea0e4e06e2b405" ], "author": { "name": "Christian Borntraeger", "email": "borntraeger@de.ibm.com", "time": "Tue May 26 15:46:09 2009 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:40 2009 +0930" }, "message": "virtio: fix id_matching for virtio drivers\n\nThis bug never appeared, since all current virtio drivers use\nVIRTIO_DEV_ANY_ID for the vendor field. If a real vendor would be used,\nthe check in virtio_id_match is wrong - it returns 0 if\nid-\u003evendor \u003d\u003d dev-\u003eid.vendor.\n\nSigned-off-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "9fa29b9df32ba4db055f3977933cd0c1b8fe67cd", "tree": "c44177d2d352f74300fe9124decc6462c8804bbe", "parents": [ "ee006b353f1ca8c9a8470b72b462beb011d62e32" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Mon May 11 18:11:45 2009 +0100" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:39 2009 +0930" }, "message": "virtio: indirect ring entries (VIRTIO_RING_F_INDIRECT_DESC)\n\nAdd a new feature flag for indirect ring entries. These are ring\nentries which point to a table of buffer descriptors.\n\nThe idea here is to increase the ring capacity by allowing a larger\neffective ring size whereby the ring size dictates the number of\nrequests that may be outstanding, rather than the size of those\nrequests.\n\nThis should be most effective in the case of block I/O where we can\npotentially benefit by concurrently dispatching a large number of\nlarge requests. Even in the simple case of single segment block\nrequests, this results in a threefold increase in ring capacity.\n\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "a92892825a122a74ddad1d408fa27132e28b05ae", "tree": "c8dad18eb7842a10a33cdf632b5e0a0aa8fdf778", "parents": [ "82af8ce84ed65d2fb6d8c017d3f2bbbf161061fb" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:37 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:38 2009 +0930" }, "message": "virtio: expose features in sysfs\n\nEach device negotiates feature bits; expose these in sysfs to help\ndiagnostics and debugging.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "82af8ce84ed65d2fb6d8c017d3f2bbbf161061fb", "tree": "aed59c86724c7d3f565d1c60b38d25a7bf20985b", "parents": [ "77cf524654a886e0fbbf03b16b44f048deef7b0c" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Thu May 14 13:55:41 2009 +0300" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:37 2009 +0930" }, "message": "virtio_pci: optional MSI-X support\n\nThis implements optional MSI-X support in virtio_pci.\nMSI-X is used whenever the host supports at least 2 MSI-X\nvectors: 1 for configuration changes and 1 for virtqueues.\nPer-virtqueue vectors are allocated if enough vectors\navailable.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nAcked-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e (+ whitespace, style)\n" }, { "commit": "77cf524654a886e0fbbf03b16b44f048deef7b0c", "tree": "4b55884ceade70463c53517eae382f0412cc317c", "parents": [ "d2a7ddda9ffb1c8961abff6714b0f1eb925c120f" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Thu May 14 13:55:31 2009 +0300" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:37 2009 +0930" }, "message": "virtio_pci: split up vp_interrupt\n\nThis reorganizes virtio-pci code in vp_interrupt slightly, so that\nit\u0027s easier to add per-vq MSI support on top.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "d2a7ddda9ffb1c8961abff6714b0f1eb925c120f", "tree": "1090884fd260d042255255467367e4e6b6193e5d", "parents": [ "9499f5e7ed5224c40706f0cec6542a9916bc7606" ], "author": { "name": "Michael S. Tsirkin", "email": "mst@redhat.com", "time": "Fri Jun 12 22:16:36 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:36 2009 +0930" }, "message": "virtio: find_vqs/del_vqs virtio operations\n\nThis replaces find_vq/del_vq with find_vqs/del_vqs virtio operations,\nand updates all drivers. This is needed for MSI support, because MSI\nneeds to know the total number of vectors upfront.\n\nSigned-off-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e (+ lguest/9p compile fixes)\n" }, { "commit": "9499f5e7ed5224c40706f0cec6542a9916bc7606", "tree": "3e4e1b36d3d549ea356e88e6e44359a887c6ee01", "parents": [ "ef688e151c00e5d529703be9a04fd506df8bc54e" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:35 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:36 2009 +0930" }, "message": "virtio: add names to virtqueue struct, mapping from devices to queues.\n\nAdd a linked list of all virtqueues for a virtio device: this helps for\ndebugging and is also needed for upcoming interface change.\n\nAlso, add a \"name\" field for clearer debug messages.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "ef688e151c00e5d529703be9a04fd506df8bc54e", "tree": "841d0b0bcd8a5f72f8d8bdc455672d9a6e7468ca", "parents": [ "20f77f5654042cf484d8964b618faf9d620f639b" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:35 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jun 12 22:16:35 2009 +0930" }, "message": "virtio: meet virtio spec by finalizing features before using device\n\nVirtio devices are supposed to negotiate features before they start using\nthe device, but the current code doesn\u0027t do this. This is because the\ndriver\u0027s probe() function invariably has to add buffers to a virtqueue,\nor probe the disk (virtio_blk).\n\nThis currently doesn\u0027t matter since no existing backend is strict about\nthe feature negotiation. But it\u0027s possible to imagine a future feature\nwhich completely changes how a device operates: in this case, we\u0027d need\nto acknowledge it before using the device.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "84a139a985300901dfad99bd93c7345d180af860", "tree": "62a8921435f17551db2f26c8f0755f4ab508f537", "parents": [ "a489f0b555b753f9df8ddc24c7e74f657ef7ee7b" ], "author": { "name": "Marcelo Tosatti", "email": "mtosatti@redhat.com", "time": "Thu Apr 16 21:14:04 2009 -0300" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Sun Apr 19 23:14:01 2009 +0930" }, "message": "virtio: fix suspend when using virtio_balloon\n\nBreak out of wait_event_interruptible() if freezing has been requested,\nin the vballoon thread. Without this change vballoon refuses to stop and\nthe system can\u0027t suspend.\n\nSigned-off-by: Marcelo Tosatti \u003cmtosatti@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: stable@kernel.org\n" }, { "commit": "c5f841f1780dad7efb7eca092f60742d47f47d25", "tree": "6a27622d96f5327408c75b914c3a54120dc1d866", "parents": [ "3a35ce7dcefe9e80a00603a195269fbaf6e7d901" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 30 21:55:22 2009 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 30 21:55:23 2009 +1030" }, "message": "virtio: more neatening of virtio_ring macros.\n\nImpact: cleanup\n\nRoel Kluin drew attention to these macros with his patch: here I\nneaten them a little further:\n1) Add a comment on what START_USE and END_USE are checking,\n2) Brackets around _vq in BAD_RING,\n3) Neaten formatting for START_USE so it\u0027s less than 80 cols.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "3a35ce7dcefe9e80a00603a195269fbaf6e7d901", "tree": "23d814e318c36d7a9fe601be2f535da45bceec56", "parents": [ "0d34fb8e93ceba7b6dad0062dbb4a0813bacd75b" ], "author": { "name": "Roel Kluin", "email": "roel.kluin@gmail.com", "time": "Thu Jan 22 16:42:57 2009 +0100" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 30 21:55:22 2009 +1030" }, "message": "virtio: fix BAD_RING, START_US and END_USE macros\n\nImpact: cleanup\n\nfix BAD_RING, START_US and END_USE macros\n\nWhen these macros aren\u0027t called with a variable named vq as first\nargument, this would result in a build failure.\n\nSigned-off-by: Roel Kluin \u003croel.kluin@gmail.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "3fff0179e33cd7d0a688dab65700c46ad089e934", "tree": "34c30ceedb3c5ae5cb6a5fb8d1c62aed7c845054", "parents": [ "720eba31f47aeade8ec130ca7f4353223c49170f" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Tue Feb 03 13:33:53 2009 +1030" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Feb 02 19:17:56 2009 -0800" }, "message": "virtio-pci: do not oops on config change if driver not loaded\n\nThe host really shouldn\u0027t be notifying us of config changes\nbefore the device status is VIRTIO_CONFIG_S_DRIVER or\nVIRTIO_CONFIG_S_DRIVER_OK.\n\nHowever, if we do happen to be interrupted while we\u0027re not\nattached to a driver, we really shouldn\u0027t oops. Prevent\nthis simply by checking that device-\u003edriver is non-NULL\nbefore trying to notify the driver of config changes.\n\nProblem observed by doing a \"set_link virtio.0 down\" with\nQEMU before the net driver had been loaded.\n\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "63d12556703f17817a46e140704360b29b851bad", "tree": "be9741e515a075010e2fbb217ea0c41b85613ca2", "parents": [ "0aa0dc41bfd993491c2344870eee7a3b218551fb" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Mon Dec 15 12:58:27 2008 +0000" }, "committer": { "name": "Greg Kroah-Hartman", "email": "gregkh@suse.de", "time": "Tue Jan 06 10:44:34 2009 -0800" }, "message": "virtio: do not statically allocate root device\n\nWe shouldn\u0027t be statically allocating the root device object,\nso dynamically allocate it using root_device_register()\ninstead.\n\nAlso avoids this warning from \u0027rmmod virtio_pci\u0027:\n\n Device \u0027virtio-pci\u0027 does not have a release() function, it is broken and must be fixed\n\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nCc: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nAcked-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nSigned-off-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\n\n" }, { "commit": "29f9f12ec737af62835124e4a8bdb9de631f04dd", "tree": "2323188170fb6462d45f34e4dfd2f34120057a76", "parents": [ "b194aee95622f649311f8e53418a17e210ff6827" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Wed Dec 10 17:45:34 2008 +0000" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:26:10 2008 +1030" }, "message": "virtio: add PCI device release() function\n\nAdd a release() function for virtio_pci devices so as to avoid:\n\n Device \u0027virtio0\u0027 does not have a release() function, it is broken and must be fixed\n\nMove the code to free the resources associated with the device\nfrom virtio_pci_remove() into this new function. virtio_pci_remove()\nnow merely unregisters the device which should cause the final\nref to be dropped and virtio_pci_release_dev() to be called.\n\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nReported-by: Michael Tokarev \u003cmjt@tls.msk.ru\u003e\nCc: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "1b4aa2faeca1b9922033daf2475b6fc13b0ffea6", "tree": "b94e3f4ec45151e51ad1b3dfecabc5391684c725", "parents": [ "87c7d57c17ade5024d95b6ca0da249da49b0672a" ], "author": { "name": "Hollis Blanchard", "email": "hollisb@us.ibm.com", "time": "Thu Nov 13 15:48:33 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:26:04 2008 +1030" }, "message": "virtio: avoid implicit use of Linux page size in balloon interface\n\nMake the balloon interface always use 4K pages, and convert Linux pfns if\nnecessary. This patch assumes that Linux\u0027s PAGE_SHIFT will never be less than\n12.\n\nSigned-off-by: Hollis Blanchard \u003chollisb@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e (modified)\n" }, { "commit": "87c7d57c17ade5024d95b6ca0da249da49b0672a", "tree": "2ffeaae5c2e5ed9a72efec6466678fb797c4cdfa", "parents": [ "db40598863e8cbbd11053ad3c8bae89000f603f9" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:26:03 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:26:03 2008 +1030" }, "message": "virtio: hand virtio ring alignment as argument to vring_new_virtqueue\n\nThis allows each virtio user to hand in the alignment appropriate to\ntheir virtio_ring structures.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nAcked-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\n" }, { "commit": "498af14783935af487d17dbee4ac451783cbc2b7", "tree": "db7a61d88f8eee05e65ac832f571440640299869", "parents": [ "5f0d1d7f2286c8a02dab69f5f0bd51681fab161e" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:25:57 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:25:58 2008 +1030" }, "message": "virtio: Don\u0027t use PAGE_SIZE for vring alignment in virtio_pci.\n\nThat doesn\u0027t work for non-4k guests which are now appearing.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "480daab42c4dd74b3c07031ddf9031251c530c77", "tree": "c491b040071be3867aa42942a80435866fbc85cf", "parents": [ "99e0b6c8e3f30602383bcfe3f574537a02ee2bea" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:25:56 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:25:57 2008 +1030" }, "message": "virtio: Don\u0027t use PAGE_SIZE in virtio_pci.c\n\nThe virtio PCI devices don\u0027t depend on the guest page size. This matters\nnow PowerPC virtio is gaining ground (they like 64k pages).\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "99e0b6c8e3f30602383bcfe3f574537a02ee2bea", "tree": "83adb7fe6cd869df901c448602872554ed7edbdf", "parents": [ "13b1eb333beed018140be3d0b77bf34000125a34" ], "author": { "name": "Kay Sievers", "email": "kay.sievers@vrfy.org", "time": "Tue Dec 30 09:25:56 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:25:56 2008 +1030" }, "message": "virtio: struct device - replace bus_id with dev_name(), dev_set_name()\n\nThis patch is part of a larger patch series which will remove\nthe \"char bus_id[20]\" name string from struct device. The device\nname is managed in the kobject anyway, and without any size\nlimitation, and just needlessly copied into \"struct device\".\n\nTo set and read the device name dev_name(dev) and dev_set_name(dev)\nmust be used. If your code uses static kobjects, which it shouldn\u0027t\ndo, \"const char *init_name\" can be used to statically provide the\nname the registered device should have. At registration time, the\ninit_name field is cleared, to enforce the use of dev_name(dev) to\naccess the device name at a later time.\n\nWe need to get rid of all occurrences of bus_id in the entire tree\nto be able to enable the new interface. Please apply this patch,\nand possibly convert any remaining remaining occurrences of bus_id.\n\nWe want to submit a patch to -next, which will remove bus_id from\n\"struct device\", to find the remaining pieces to convert, and finally\nswitch over to the new api, which will remove the 20 bytes array\nand does no longer have a size limitation.\n\nAcked-by: Greg Kroah-Hartman \u003cgregkh@suse.de\u003e\nSigned-off-by: Kay Sievers \u003ckay.sievers@vrfy.org\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "13b1eb333beed018140be3d0b77bf34000125a34", "tree": "26d84c1c955ad87687266433e8e54fa6f88bd150", "parents": [ "3c92ec8ae91ecf59d88c798301833d7cf83f2179" ], "author": { "name": "Hollis Blanchard", "email": "hollisb@us.ibm.com", "time": "Tue Dec 02 16:24:40 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Dec 30 09:25:56 2008 +1030" }, "message": "virtio-pci queue allocation not page-aligned\n\nkzalloc() does not guarantee page alignment, and in fact this broke when\nI enabled CONFIG_SLUB_DEBUG_ON.\n\n(Thanks to Anthony Liguori for spotting the missing kfree sub)\n\nSigned-off-by: Hollis Blanchard \u003chollisb@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e (fixed kfree)\nTested-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\n" }, { "commit": "532a6086e35fa3b5761e68af36d4e42a550eba15", "tree": "c1d8ab4df9db25e0efd89dab64dda98f84df1fa6", "parents": [ "83097aca8567a0bd593534853b71fe0fa9a75d69" ], "author": { "name": "Anthony Liguori", "email": "aliguori@us.ibm.com", "time": "Mon Aug 18 17:15:31 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Aug 26 00:19:25 2008 +1000" }, "message": "virtio_balloon: fix towards_target when deflating balloon\n\nBoth v and vb-\u003enum_pages are u32 and unsigned int respectively. If v is less\nthan vb-\u003enum_pages (and it is, when deflating the balloon), the result is a\nvery large 32-bit number. Since we\u0027re returning a s64, instead of getting the\nsame negative number we desire, we get a very large positive number.\n\nThis handles the case where v \u003c vb-\u003enum_pages and ensures we get a small,\nnegative, s64 as the result.\n\nRusty: please push this for 2.6.27-rc4. It\u0027s probably appropriate for the\nstable tree too as it will cause an unexpected OOM when ballooning.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e (simplified)\n" }, { "commit": "e34f87256794b87e7f4a8f1812538be7b7b5214c", "tree": "b91ffb6df4ac13ace269bef60ae16a1048b1fa73", "parents": [ "c624896e488ba2bff5ae497782cfb265c8b00646" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:13 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:14 2008 +1000" }, "message": "virtio: Add transport feature handling stub for virtio_ring.\n\nTo prepare for virtio_ring transport feature bits, hook in a call in\nall the users to manipulate them. This currently just clears all the\nbits, since it doesn\u0027t understand any features.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "c624896e488ba2bff5ae497782cfb265c8b00646", "tree": "714145903ff8c889e9544e8366b2ee2a62234c0c", "parents": [ "dd7c7bc46211785a1aa7d70feb15830f62682b3c" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:07 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:12 2008 +1000" }, "message": "virtio: Rename set_features to finalize_features\n\nRather than explicitly handing the features to the lower-level, we just\nhand the virtio_device and have it set the features. This make it clear\nthat it has the chance to manipulate the features of the device at this\npoint (and that all feature negotiation is already done).\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "dd7c7bc46211785a1aa7d70feb15830f62682b3c", "tree": "3c615eac390c11339c652947e744d12cb82b4d36", "parents": [ "faeba830b086bc9e58748869054e994cb09693cd" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:07 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:07 2008 +1000" }, "message": "virtio: Formally reserve bits 28-31 to be \u0027transport\u0027 features.\n\nWe assign feature bits as required, but it makes sense to reserve some\nfor the particular transport, rather than the particular device.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "e962fa660d391fc9b90988e6538c94c858c099f9", "tree": "2e498cb10c9a416192c769bec8119455afcb18ba", "parents": [ "44653eae1407f79dff6f52fcf594ae84cb165ec4" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Fri Jun 13 13:46:40 2008 +0100" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:05 2008 +1000" }, "message": "virtio: Use bus_type probe and remove methods\n\nHook up to the probe() and remove() methods in bus_type\nrather than device_driver. The latter has been preferred\nsince 2.6.16.\n\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "44653eae1407f79dff6f52fcf594ae84cb165ec4", "tree": "9c8f356f6c294678a05dbc983e60b8b5d864b7bc", "parents": [ "674bfc23c585b34c42263d73fb51710d49762a23" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:04 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Jul 25 12:06:04 2008 +1000" }, "message": "virtio: don\u0027t always force a notification when ring is full\n\nWe force notification when the ring is full, even if the host has\nindicated it doesn\u0027t want to know. This seemed like a good idea at\nthe time: if we fill the transmit ring, we should tell the host\nimmediately.\n\nUnfortunately this logic also applies to the receiving ring, which is\nrefilled constantly. We should introduce real notification thesholds\nto replace this logic. Meanwhile, removing the logic altogether breaks\nthe heuristics which KVM uses, so we use a hack: only notify if there are\noutgoing parts of the new buffer.\n\nHere are the number of exits with lguest\u0027s crappy network implementation:\nBefore:\n\tnetwork xmit 7859051 recv 236420\nAfter:\n\tnetwork xmit 7858610 recv 118136\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "b92dea67cc66970cda6b5b11895d08e35b4618e7", "tree": "556e842834c7316f57f3f27fea6b8622d74fde8e", "parents": [ "0269c5c6d9a9de22715ecda589730547435cd3e8" ], "author": { "name": "Mark McLoughlin", "email": "markmc@redhat.com", "time": "Sun Jun 15 23:20:50 2008 +1000" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 15 13:46:16 2008 -0700" }, "message": "virtio: Complete feature negotation before updating status\n\nlguest (in rusty\u0027s use-tun-ringfd patch) assumes that the\nguest has updated its feature bits before setting its status\nto VIRTIO_CONFIG_S_DRIVER_OK.\n\nThat\u0027s pretty reasonable, so let\u0027s make it so.\n\nSigned-off-by: Mark McLoughlin \u003cmarkmc@redhat.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b4f68be6c5d507afdcd74f5be3df0b1209cda503", "tree": "85c0771058ff08c5dab5eedbf3395959dbafc878", "parents": [ "7757f09c70af87887dfc195e6d6ddd54f5cc7c39" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:45 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:46 2008 +1000" }, "message": "virtio: force callback on empty.\n\nvirtio allows drivers to suppress callbacks (ie. interrupts) for\nefficiency (no locking, it\u0027s just an optimization).\n\nThere\u0027s a similar mechanism for the host to suppress notifications\ncoming from the guest: in that case, we ignore the suppression if the\nring is completely full.\n\nIt turns out that life is simpler if the host similarly ignores\ncallback suppression when the ring is completely empty: the network\ndriver wants to free up old packets in a timely manner, and otherwise\nhas to use a timer to poll.\n\nWe have to remove the code which ignores interrupts when the driver\nhas disabled them (again, it had no locking and hence was unreliable\nanyway).\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "52a3a05f3ab82655ffa4c9bf6835565c98a3c2e5", "tree": "f4257904896d265059d4a9f330e8bf5d3181c6f9", "parents": [ "f7f510ec195781c857ab76366a3e1c59e1caae42" ], "author": { "name": "Christian Borntraeger", "email": "borntraeger@de.ibm.com", "time": "Mon May 26 11:29:27 2008 +0200" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:45 2008 +1000" }, "message": "virtio_net: another race with virtio_net and enable_cb\n\nHello Rusty,\n\nseems that we still have a problem with virtio_net and the enable_cb callback.\nDuring a long running network stress tests with virtio and got the following\noops:\n\n------------[ cut here ]------------\nkernel BUG at drivers/virtio/virtio_ring.c:230!\nillegal operation: 0001 [#1] SMP\nModules linked in:\nCPU: 0 Not tainted 2.6.26-rc2-kvm-00436-gc94c08b-dirty #34\nProcess netserver (pid: 2582, task: 000000000fbc4c68, ksp: 000000000f42b990)\nKrnl PSW : 0704c00180000000 00000000002d0ec8 (vring_enable_cb+0x1c/0x60)\n R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:0 PM:0 EA:3\nKrnl GPRS: 0000000000000000 0000000000000000 000000000ef3d000 0000000010009800\n 0000000000000000 0000000000419ce0 0000000000000080 000000000000007b\n 000000000adb5538 000000000ef40900 000000000ef40000 000000000ef40920\n 0000000000000000 0000000000000005 000000000029c1b0 000000000fea7d18\nKrnl Code: 00000000002d0ebc: a7110001 tmll %r1,1\n 00000000002d0ec0: a7740004 brc 7,2d0ec8\n 00000000002d0ec4: a7f40001 brc 15,2d0ec6\n \u003e00000000002d0ec8: a517fffe nill %r1,65534\n 00000000002d0ecc: 40103000 sth %r1,0(%r3)\n 00000000002d0ed0: 07f0 bcr 15,%r0\n 00000000002d0ed2: e31020380004 lg %r1,56(%r2)\n 00000000002d0ed8: a7480000 lhi %r4,0\nCall Trace:\n([\u003c000000000029c0fc\u003e] virtnet_poll+0x290/0x3b8)\n [\u003c0000000000333fb8\u003e] net_rx_action+0x9c/0x1b8\n [\u003c00000000001394bc\u003e] __do_softirq+0x74/0x108\n [\u003c000000000010d16a\u003e] do_softirq+0x92/0xac\n [\u003c0000000000139826\u003e] irq_exit+0x72/0xc8\n [\u003c000000000010a7b6\u003e] do_extint+0xe2/0x104\n [\u003c0000000000110508\u003e] ext_no_vtime+0x16/0x1a\nLast Breaking-Event-Address:\n [\u003c00000000002d0ec4\u003e] vring_enable_cb+0x18/0x60\n\nI looked into the virtio_net code for some time and I think the following\nscenario happened. Please look at virtnet_poll:\n[...]\n /* Out of packets? */\n if (received \u003c budget) {\n netif_rx_complete(vi-\u003edev, napi);\n if (unlikely(!vi-\u003ervq-\u003evq_ops-\u003eenable_cb(vi-\u003ervq))\n \u0026\u0026 napi_schedule_prep(napi)) {\n vi-\u003ervq-\u003evq_ops-\u003edisable_cb(vi-\u003ervq);\n __netif_rx_schedule(vi-\u003edev, napi);\n goto again;\n }\n }\n\nIf an interrupt arrives after netif_rx_complete, a second poll routine can run\non a different cpu. The second check for napi_schedule_prep would prevent any\nharm in the network stack, but we have called enable_cb possibly after the\ndisable_cb in skb_recv_done.\n\nstatic void skb_recv_done(struct virtqueue *rvq)\n{\n struct virtnet_info *vi \u003d rvq-\u003evdev-\u003epriv;\n /* Schedule NAPI, Suppress further interrupts if successful. */\n if (netif_rx_schedule_prep(vi-\u003edev, \u0026vi-\u003enapi)) {\n rvq-\u003evq_ops-\u003edisable_cb(rvq);\n __netif_rx_schedule(vi-\u003edev, \u0026vi-\u003enapi);\n }\n}\n\nThat means that the second poll routine runs with interrupts enabled, which is\nok, since we can handle additional interrupts. The problem is now that the\nsecond poll routine might also call enable_cb, triggering the BUG.\n\nThe only solution I can come up with, is to remove the BUG statement in\nenable_cb - similar to disable_cb. Opinions or better ideas where the oops\ncould come from?\n\nSigned-off-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "b769f579081943f14e0ff03b7b0bd3a11cf14625", "tree": "026d89b6d937af43af4a692bcf547e032f0c58cc", "parents": [ "5610bd1524332fe7d651eb56cc780e32763a2ac3" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:42 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:42 2008 +1000" }, "message": "virtio: set device index in common code.\n\nAnthony Liguori points out that three different transports use the virtio code,\nbut each one keeps its own counter to set the virtio_device\u0027s index field. In\ntheory (though not in current practice) this means that names could be\nduplicated, and that risk grows as more transports are created.\n\nSo we move the selection of the unique virtio_device.index into the common code\nin virtio.c, which has the side-benefit of removing duplicate code.\n\nThe only complexity is that lguest and S/390 use the index to uniquely identify\nthe device in case of catastrophic failure before register_virtio_device() is\ncalled: now we use the offset within the descriptor page as a unique identifier\nfor the printks.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nCc: Martin Schwidefsky \u003cschwidefsky@de.ibm.com\u003e\nCc: Carsten Otte \u003ccotte@de.ibm.com\u003e\nCc: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\nCc: Chris Lalancette \u003cclalance@redhat.com\u003e\nCc: Anthony Liguori \u003canthony@codemonkey.ws\u003e\n" }, { "commit": "5610bd1524332fe7d651eb56cc780e32763a2ac3", "tree": "551e8ea4cbbf087f3fe23f10eebb075466369228", "parents": [ "2ad3cfbac58d0a6c6e65aafd9e0e757ca3d35292" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:42 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:42 2008 +1000" }, "message": "virtio: virtio_pci should not set bus_id.\n\nThe common virtio code sets the bus_id, overriding anything virtio_pci\nsets anyway.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nCc: Martin Schwidefsky \u003cschwidefsky@de.ibm.com\u003e\nCc: Carsten Otte \u003ccotte@de.ibm.com\u003e\nCc: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\nCc: Chris Lalancette \u003cclalance@redhat.com\u003e\nCc: Anthony Liguori \u003canthony@codemonkey.ws\u003e\n" }, { "commit": "2ad3cfbac58d0a6c6e65aafd9e0e757ca3d35292", "tree": "fd17937a5873c078487a02adc5dbdc8a175b0796", "parents": [ "ac9d463afb1ca2434335351f3b7d9e4c8f8470e9" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:41 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 30 15:09:42 2008 +1000" }, "message": "virtio: bus_id for devices should contain \u0027virtio\u0027\n\nChris Lalancette \u003cclalance@redhat.com\u003e points out that virtio.c sets all device\nnames to \u00270\u0027, \u00271\u0027, etc, which looks silly in /proc/interrupts. We change this\nfrom \u0027%d\u0027 to \u0027virtio%d\u0027.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nCc: Martin Schwidefsky \u003cschwidefsky@de.ibm.com\u003e\nCc: Carsten Otte \u003ccotte@de.ibm.com\u003e\nCc: Heiko Carstens \u003cheiko.carstens@de.ibm.com\u003e\nCc: Chris Lalancette \u003cclalance@redhat.com\u003e\nCc: Anthony Liguori \u003canthony@codemonkey.ws\u003e\n" }, { "commit": "c45a6816c19dee67b8f725e6646d428901a6dc24", "tree": "096e3263fd14e140685bcc3082394ff15f5aeddb", "parents": [ "72e61eb40b55dd57031ec5971e810649f82b0259" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:50 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:50 2008 +1000" }, "message": "virtio: explicit advertisement of driver features\n\nA recent proposed feature addition to the virtio block driver revealed\nsome flaws in the API: in particular, we assume that feature\nnegotiation is complete once a driver\u0027s probe function returns.\n\nThere is nothing in the API to require this, however, and even I\ndidn\u0027t notice when it was violated.\n\nSo instead, we require the driver to specify what features it supports\nin a table, we can then move the feature negotiation into the virtio\ncore. The intersection of device and driver features are presented in\na new \u0027features\u0027 bitmap in the struct virtio_device.\n\nNote that this highlights the difference between Linux unsigned-long\nbitmaps where each unsigned long is in native endian, and a\nstraight-forward little-endian array of bytes.\n\nDrivers can still remove feature bits in their probe routine if they\nreally have to.\n\nAPI changes:\n- dev-\u003econfig-\u003efeature() no longer gets and acks a feature.\n- drivers should advertise their features in the \u0027feature_table\u0027 field\n- use virtio_has_feature() for extra sanity when checking feature bits\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "72e61eb40b55dd57031ec5971e810649f82b0259", "tree": "66a836c7799b21156d4fc87f42e5817d7d95535b", "parents": [ "5539ae9613587e4a4eec42d420b8bdd9ff552a65" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:49 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:50 2008 +1000" }, "message": "virtio: change config to guest endian.\n\nA recent proposed feature addition to the virtio block driver revealed\nsome flaws in the API, in particular how easy it is to break big\nendian machines.\n\nThe virtio config space was originally chosen to be little-endian,\nbecause we thought the config might be part of the PCI config space\nfor virtio_pci. It\u0027s actually a separate mmio region, so that\nargument holds little water; as only x86 is currently using the virtio\nmechanism, we can change this (but must do so now, before the\nimpending s390 merge).\n\nAPI changes:\n- __virtio_config_val() just becomes a striaght vdev-\u003econfig_get() call.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "597d56e4b51fc3385e097e52d6e92bf596ff21ec", "tree": "7e5db5e9491ac0c24c1f2bea979dc9d957d7b537", "parents": [ "655aa31f028c4498e8896576571ee1ea68dd26e0" ], "author": { "name": "Harvey Harrison", "email": "harvey.harrison@gmail.com", "time": "Mon Mar 31 17:53:55 2008 -0700" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:44 2008 +1000" }, "message": "virtio: fix sparse return void-valued expression warnings\n\ndrivers/virtio/virtio_pci.c:148:2: warning: returning void-valued expression\ndrivers/virtio/virtio_pci.c:155:2: warning: returning void-valued expression\n\nSigned-off-by: Harvey Harrison \u003charvey.harrison@gmail.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "5ef827526fc01820a7a80827802e9fad3f34f937", "tree": "c41345c6de6ad5b4b4d7071aa46a9e27a773247c", "parents": [ "886c35fbcf6fb2eee15687efc2d64d99b6ad9a4a" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:43 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri May 02 21:50:43 2008 +1000" }, "message": "virtio: ignore corrupted virtqueues rather than spinning.\n\nA corrupt virtqueue (caused by the other end screwing up) can have\nstrange results such as a driver spinning: just bail when we try to\nget a buffer from a known-broken queue.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "2557a933b795c1988c721ebb871cd735128bb9cb", "tree": "96581bd5acac44bbe0399aa445197509d9ae60ff", "parents": [ "4cac04dd63fa3b202ee313ed1afbbd135ab887ee" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Apr 07 14:30:28 2008 +1000" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Apr 07 13:14:22 2008 -0700" }, "message": "virtio: remove overzealous BUG_ON.\n\nThe \u0027disable_cb\u0027 callback is designed as an optimization to tell the host\nwe don\u0027t need callbacks now. As it is not reliable, the debug check is\noverzealous: it can happen on two CPUs at the same time. Document this.\n\nEven if it were reliable, the virtio_net driver doesn\u0027t disable\ncallbacks on transmit so the START_USE/END_USE debugging reentrance\nprotection can be easily tripped even on UP.\n\nThanks to Balaji Rao for the bug report and testing.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCC: Balaji Rao \u003cbalajirrao@gmail.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "97968358ab98c2d5fc5c3071ba70bd6a5faecbb8", "tree": "09201cba8cdff97f11f14aa3ea0517942ccb2948", "parents": [ "b2ddb9019ea13fb7b62d8e45adcc468376af0de7" ], "author": { "name": "Al Viro", "email": "viro@ftp.linux.org.uk", "time": "Sat Mar 29 03:09:48 2008 +0000" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Mar 30 14:20:23 2008 -0700" }, "message": "virtio_pci iomem annotations\n\nSigned-off-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "bd6c26900bae19bd51abedfc9a8a281afbba3a06", "tree": "cb2058d113d81fa4b168b8675f000f8aecbf8b26", "parents": [ "3085354de635179d70c240e6d942bcbd1d93056c" ], "author": { "name": "Anthony Liguori", "email": "aliguori@us.ibm.com", "time": "Wed Mar 19 20:35:04 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Mar 28 11:05:51 2008 +1100" }, "message": "virtio_pci: unregister virtio device at device remove\n\nMake sure to call unregister_virtio_device() when a virtio device is removed.\nOtherwise, virtio_pci.ko cannot be rmmod\u0027d.\n\nThis was spotted by Marcelo Tosatti.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "4265f161b6bb7b31163671329b1142b9023bf4e3", "tree": "1494768aaf38407ca8f1336f6ae80301b27135bb", "parents": [ "da74e89d40995600b3b07ac500084920247687ce" ], "author": { "name": "Christian Borntraeger", "email": "borntraeger@de.ibm.com", "time": "Fri Mar 14 14:17:05 2008 +0100" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 17 22:58:21 2008 +1100" }, "message": "virtio: fix race in enable_cb\n\nThere is a race in virtio_net, dealing with disabling/enabling the callback.\nI saw the following oops:\n\nkernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:218!\nillegal operation: 0001 [#1] SMP\nModules linked in: sunrpc dm_mod\nCPU: 2 Not tainted 2.6.25-rc1zlive-host-10623-gd358142-dirty #99\nProcess swapper (pid: 0, task: 000000000f85a610, ksp: 000000000f873c60)\nKrnl PSW : 0404300180000000 00000000002b81a6 (vring_disable_cb+0x16/0x20)\n R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3\nKrnl GPRS: 0000000000000001 0000000000000001 0000000010005800 0000000000000001\n 000000000f3a0900 000000000f85a610 0000000000000000 0000000000000000\n 0000000000000000 000000000f870000 0000000000000000 0000000000001237\n 000000000f3a0920 000000000010ff74 00000000002846f6 000000000fa0bcd8\nKrnl Code: 00000000002b819a: a7110001 tmll %r1,1\n 00000000002b819e: a7840004 brc 8,2b81a6\n 00000000002b81a2: a7f40001 brc 15,2b81a4\n \u003e00000000002b81a6: a51b0001 oill %r1,1\n 00000000002b81aa: 40102000 sth %r1,0(%r2)\n 00000000002b81ae: 07fe bcr 15,%r14\n 00000000002b81b0: eb7ff0380024 stmg %r7,%r15,56(%r15)\n 00000000002b81b6: a7f13e00 tmll %r15,15872\nCall Trace:\n([\u003c000000000fa0bcd0\u003e] 0xfa0bcd0)\n [\u003c00000000002b8350\u003e] vring_interrupt+0x5c/0x6c\n [\u003c000000000010ab08\u003e] do_extint+0xb8/0xf0\n [\u003c0000000000110716\u003e] ext_no_vtime+0x16/0x1a\n [\u003c0000000000107e72\u003e] cpu_idle+0x1c2/0x1e0\n\nThe problem can be triggered with a high amount of host-\u003eguest traffic.\nI think its the following race:\n\npoll says netif_rx_complete\npoll calls enable_cb\nenable_cb opens the interrupt mask\na new packet comes, an interrupt is triggered----\\\nenable_cb sees that there is more work |\nenable_cb disables the interrupt |\n . V\n . interrupt is delivered\n . skb_recv_done does atomic napi test, ok\n some waiting disable_cb is called-\u003echeck fails-\u003ebang!\n .\npoll would do napi check\npoll would do disable_cb\n\nThe fix is to let enable_cb not disable the interrupt again, but expect the\ncaller to do the cleanup if it returns false. In that case, the interrupt is\nonly disabled, if the napi test_set_bit was successful.\n\nSigned-off-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e (cleaned up doco)\n" }, { "commit": "bdc1681cdf1ab6a65fa935a2b3f8fc63b20c54ea", "tree": "17910bdca2cee81b0d4a3ee877aec82db0017a09", "parents": [ "c483934670d31e064e18967bb679c1079b54ea72" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 17 22:58:15 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 17 22:58:19 2008 +1100" }, "message": "virtio: handle \u003e 2 billion page balloon targets\n\nIf the host asks for a huge target towards_target() can overflow, and\nwe up oops as we try to release more pages than we have. The simple\nfix is to use a 64-bit value.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "27ebe308af7c14aed0711e25817c6a1235bd16ff", "tree": "dafd3caaa1591a60a7b79f7027c00bcae30f4a6e", "parents": [ "a978b30af3bab0dd9af9350eeda25e76123fa28e" ], "author": { "name": "Anthony Liguori", "email": "aliguori@us.ibm.com", "time": "Sun Mar 02 16:37:48 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Mar 17 22:58:13 2008 +1100" }, "message": "virtio: Use spin_lock_irqsave/restore for virtio-pci\n\nvirtio-pci acquires its spin lock in an interrupt context so it\u0027s necessary\nto use spin_lock_irqsave/restore variants. This patch fixes guest SMP when\nusing virtio devices in KVM.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "6659a0f0bb7481d0f94ca3f203a4e1e406a9694d", "tree": "6803f9ad7ceb8556e8a07a8ba2e2d1a61e7283f1", "parents": [ "bd1939de9061dbc5cac44ffb4425aaf4c9b894f1" ], "author": { "name": "Johann Felix Soden", "email": "johfel@users.sourceforge.net", "time": "Wed Feb 06 01:40:22 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Wed Feb 06 10:41:21 2008 -0800" }, "message": "virtio: add missing #include \u003clinux/delay.h\u003e\n\nInclude linux/delay.h to fix compiler error:\n\ndrivers/virtio/virtio_balloon.c: In function \u0027fill_balloon\u0027:\ndrivers/virtio/virtio_balloon.c:98: error: implicit declaration of function \u0027msleep\u0027\n\nSigned-off-by: Johann Felix Soden \u003cjohfel@users.sourceforge.net\u003e\nCc: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6b35e40767c6c1ac783330109ae8e0c09ea6bc82", "tree": "e820d770894cb945c3070fa6739b0149b1bfa2f9", "parents": [ "55a7c066041e7850948d29ed813f62821a9ec046" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:12 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:13 2008 +1100" }, "message": "virtio: balloon driver\n\nAfter discussions with Anthony Liguori, it seems that the virtio\nballoon can be made even simpler. Here\u0027s my attempt.\n\nThe device configuration tells the driver how much memory it should\ntake from the guest (ie. balloon size). The guest feeds the page\nnumbers it has taken via one virtqueue.\n\nA second virtqueue feeds the page numbers the driver wants back: if\nthe device has the VIRTIO_BALLOON_F_MUST_TELL_HOST bit, then this\nqueue is compulsory, otherwise it\u0027s advisory (and the guest can simply\nfault the pages back in).\n\nThis driver can be enhanced later to deflate the balloon via a\nshrinker, oom callback or we could even go for a complete set of\nin-guest regulators.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "55a7c066041e7850948d29ed813f62821a9ec046", "tree": "ea4ef20c8014d5222957cef83d4fc62b75c58fd6", "parents": [ "3343660d8c62c6b00b2f15324ef3fcb6be207bfa" ], "author": { "name": "Anthony Liguori", "email": "aliguori@us.ibm.com", "time": "Mon Jan 28 09:59:59 2008 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:12 2008 +1100" }, "message": "virtio: Use PCI revision field to indicate virtio PCI ABI version\n\nAs Avi pointed out, as we continue to massage the virtio PCI ABI, we can make\nthings a little more friendly to users by utilizing the PCI revision field to\nindicate which version of the ABI we\u0027re using. This is a hard ABI version\nand incrementing it will cause the guest driver to break.\n\nThis is the necessary changes to virtio_pci to support this.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "3343660d8c62c6b00b2f15324ef3fcb6be207bfa", "tree": "06d86446da65fc1814edad944e43aeb62f092422", "parents": [ "d50ed907dc3db5bf2dd0a05b4e199a65793a3788" ], "author": { "name": "Anthony Liguori", "email": "aliguori@us.ibm.com", "time": "Mon Nov 12 21:30:26 2007 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:11 2008 +1100" }, "message": "virtio: PCI device\n\nThis is a PCI device that implements a transport for virtio. It allows virtio\ndevices to be used by QEMU based VMMs like KVM or Xen.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "c6fd47011b4bdebad3f1513bac75fe4895e332ee", "tree": "12224cdb2594be26e02291b142ba6e1e2c1c8122", "parents": [ "15f9c8903cbdb02aee0f1bcf86a97c2e238b9a3d" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:05 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:06 2008 +1100" }, "message": "virtio: Allow virtio to be modular and used by modules\n\nThis is needed for the virtio PCI device to be compiled as a module.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "15f9c8903cbdb02aee0f1bcf86a97c2e238b9a3d", "tree": "a215ccf3f745e52ae67aed1ac822baa57326b9b6", "parents": [ "0ad07ec1fd2f3e90dfb086b04c07a83db40e9172" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:05 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:05 2008 +1100" }, "message": "virtio: Use the sg_phys convenience function.\n\nSimple cleanup.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "81a8deab1ce3816c6a89e3429e234e7d3686da94", "tree": "d62e7dcc09dddd0edc6d507e7b219c16e66d3b02", "parents": [ "6e5aa7efb27aec7e55b6463fa2c8db594c4226fa" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:04 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:04 2008 +1100" }, "message": "virtio: handle interrupts after callbacks turned off\n\nAnthony Liguori found double interrupt suppression in the virtio_net\ndriver, triggered by two skb_recv_done\u0027s in a row. This is because\nvirtio_ring\u0027s interrupt suppression is a best-effort optimization: it\ncontains no synchronization so the host can miss it and still send\ninterrupts.\n\nBut it\u0027s certainly nicer for virtio users if calling disable_cb\nactually disables callbacks, so we check for the race in the interrupt\nroutine.\n\nNote: SMP guests might require syncronization here, but since\ndisable_cb is actually called from interrupt context, there has to be\nsome form of synchronization before the next same interrupt handler is\ncalled (Linux guarantees that the same device\u0027s irq handler will never\nrun simultanously on multiple CPUs).\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "6e5aa7efb27aec7e55b6463fa2c8db594c4226fa", "tree": "060a955e711ac224136157a5410e88dcdab965af", "parents": [ "b3369c1fb410fddeb38a404316c861395f6d6ae8" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:03 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:03 2008 +1100" }, "message": "virtio: reset function\n\nA reset function solves three problems:\n\n1) It allows us to renegotiate features, eg. if we want to upgrade a\n guest driver without rebooting the guest.\n\n2) It gives us a clean way of shutting down virtqueues: after a reset,\n we know that the buffers won\u0027t be used by the host, and\n\n3) It helps the guest recover from messed-up drivers.\n\nSo we remove the -\u003eshutdown hook, and the only way we now remove\nfeature bits is via reset.\n\nWe leave it to the driver to do the reset before it deletes queues:\nthe balloon driver, for example, needs to chat to the host in its\nremove function.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "426e3e0af5d2473e67d4256fc1340b7faebd1cc7", "tree": "29912e63889e45307f41ef08116ecd895b17e21c", "parents": [ "3309daaad724dd08eb598bf9c12b7bb9daddd706" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:49:59 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:50:00 2008 +1100" }, "message": "virtio: clarify NO_NOTIFY flag usage\n\nThe other side (host) can set the NO_NOTIFY flag as an optimization,\nto say \"no need to kick me when you add things\". Make it clear that\nthis is advisory only; especially that we should always notify when\nthe ring is full.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "18445c4d501b9ab4336f66ef46b092661ddaf336", "tree": "9d23185f207d912942890cf047d1d3200806b401", "parents": [ "a586d4f6016f7139d8c26df0e6927131168d3b5b" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:49:57 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:49:58 2008 +1100" }, "message": "virtio: explicit enable_cb/disable_cb rather than callback return.\n\nIt seems that virtio_net wants to disable callbacks (interrupts) before\ncalling netif_rx_schedule(), so we can\u0027t use the return value to do so.\n\nRename \"restart\" to \"cb_enable\" and introduce \"cb_disable\" hook: callback\nnow returns void, rather than a boolean.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "a586d4f6016f7139d8c26df0e6927131168d3b5b", "tree": "1c47e1a6b6b8fb18baa42f32980f29c4ae9cbbdc", "parents": [ "f35d9d8aae08940b7fdd1bb8110619da2ece6b28" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:49:56 2008 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Feb 04 23:49:57 2008 +1100" }, "message": "virtio: simplify config mechanism.\n\nPreviously we used a type/len pair within the config space, but this\nseems overkill. We now simply define a structure which represents the\nlayout in the config space: the config space can now only be extended\nat the end.\n\nThe main driver-visible changes:\n1) We indicate what fields are present with an explicit feature bit.\n2) Virtqueues are explicitly numbered, and not in the config space.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "74b2553f1d13e60fb27063204bd5b6908a6f8494", "tree": "cd35e82d16cf190ccd95362478a598314de639ce", "parents": [ "d1c856e0f1a4c946c6329cff126548ef4288735f" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Nov 19 11:20:42 2007 -0500" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Nov 19 11:20:42 2007 +1100" }, "message": "virtio: fix module/device unloading\n\nThe virtio code never hooked through the -\u003eremove callback. Although\nnoone supports device removal at the moment, this code is already\nneeded for module unloading.\n\nThis of course also revealed bugs in virtio_blk, virtio_net and lguest\nunloading paths.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "42b36cc0ce717deeb10030141a43dede763a3ebe", "tree": "b2dc48b4f16c5dc59461ad24b027d631edda1da4", "parents": [ "1200e646ae238afc536be70257290eb33fb6e364" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Nov 12 13:39:18 2007 +1100" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Nov 12 13:59:40 2007 +1100" }, "message": "virtio: Force use of power-of-two for descriptor ring sizes\n\nThe virtio descriptor rings of size N-1 were nicely set up to be\naligned to an N-byte boundary. But as Anthony Liguori points out, the\nfree-running indices used by virtio require that the sizes be a power\nof 2, otherwise we get problems on wrap (demonstrated with lguest).\n\nSo we replace the clever \"2^n-1\" scheme with a simple \"align to page\nboundary\" scheme: this means that all virtio rings take at least two\npages, but it\u0027s safer than guessing cache alignment.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "1bc4953ed44454c7f53d0b609445d1534981ee75", "tree": "05f67206d6978261ad2569d3a0030f0299331191", "parents": [ "4d125de3a5d130054df2285e542c1491d214d3e8" ], "author": { "name": "Anthony Liguori", "email": "aliguori@us.ibm.com", "time": "Wed Nov 07 15:49:24 2007 -0600" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Nov 12 13:59:09 2007 +1100" }, "message": "virtio: Fix used_idx wrap-around\n\nThe more_used() function compares the vq-\u003evring.used-\u003eidx with last_used_idx.\nSince vq-\u003evring.used-\u003eidx is a 16-bit integer, and last_used_idx is an\nunsigned int, this results in unpredictable behavior when vq-\u003evring.used-\u003eidx\nwraps around.\n\nThis patch corrects this by changing last_used_idx to the correct type.\n\nSigned-off-by: Anthony Liguori \u003caliguori@us.ibm.com\u003e\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "0a8a69dd77ddbd4513b21363021ecde7e1025502", "tree": "ed6d8f0756835390b4c0d9a172422f2e42a65523", "parents": [ "b01d9f2863349b0e041b90c3c86a998ee0fed2b0" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Oct 22 11:03:40 2007 +1000" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Oct 23 15:49:55 2007 +1000" }, "message": "Virtio helper routines for a descriptor ringbuffer implementation\n\nThese helper routines supply most of the virtqueue_ops for hypervisors\nwhich want to use a ring for virtio. Unlike the previous lguest\nimplementation:\n\n1) The rings are variable sized (2^n-1 elements).\n2) They have an unfortunate limit of 65535 bytes per sg element.\n3) The page numbers are always 64 bit (PAE anyone?)\n4) They no longer place used[] on a separate page, just a separate\n cacheline.\n5) We do a modulo on a variable. We could be tricky if we cared.\n6) Interrupts and notifies are suppressed using flags within the rings.\n\nUsers need only get the ring pages and provide a notify hook (KVM\nwants the guest to allocate the rings, lguest does it sanely).\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Dor Laor \u003cdor.laor@qumranet.com\u003e\n" }, { "commit": "b01d9f2863349b0e041b90c3c86a998ee0fed2b0", "tree": "8789a06631a0ee502c58388d12c230eb4a2059b2", "parents": [ "31610434bc3523c0b01a10917a1185096a03c4c8" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Oct 22 11:03:39 2007 +1000" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Oct 23 15:49:55 2007 +1000" }, "message": "Module autoprobing support for virtio drivers.\n\nThis adds the logic to convert the virtio ids into module aliases, and\nincludes a modalias entry in sysfs and the env var to make probing work.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "ec3d41c4db4c21164332826ea8d812f94f2f6886", "tree": "9b947e900745cfcc4df7409c6ba5583428964215", "parents": [ "47436aa4ad054c1c7c8231618e86ebd9305308dc" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Mon Oct 22 11:03:36 2007 +1000" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Oct 23 15:49:54 2007 +1000" }, "message": "Virtio interface\n\nThis attempts to implement a \"virtual I/O\" layer which should allow\ncommon drivers to be efficiently used across most virtual I/O\nmechanisms. It will no-doubt need further enhancement.\n\nThe virtio drivers add buffers to virtio queues; as the buffers are consumed\nthe driver \"interrupt\" callbacks are invoked.\n\nThere is also a generic implementation of config space which drivers can query\nto get setup information from the host.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Dor Laor \u003cdor.laor@qumranet.com\u003e\nCc: Arnd Bergmann \u003carnd@arndb.de\u003e\n" } ] }