)]}' { "log": [ { "commit": "59534f7298c5e28aaa64e6ed550e247f64ee72ae", "tree": "b9fef7756abf897d9e1b10950cdf10bf6dfe5cb7", "parents": [ "ac3ee84c604502240122c47b52f0542ec8774f15", "b486787ee4797d6e42a9bd3a6f079385ad0f4472" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 21 11:14:52 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 21 11:14:52 2010 -0700" }, "message": "Merge branch \u0027drm-for-2.6.35\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6\n\n* \u0027drm-for-2.6.35\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (207 commits)\n drm/radeon/kms/pm/r600: select the mid clock mode for single head low profile\n drm/radeon: fix power supply kconfig interaction.\n drm/radeon/kms: record object that have been list reserved\n drm/radeon: AGP memory is only I/O if the aperture can be mapped by the CPU.\n drm/radeon/kms: don\u0027t default display priority to high on rs4xx\n drm/edid: fix typo in 1600x1200@75 mode\n drm/nouveau: fix i2c-related init table handlers\n drm/nouveau: support init table i2c device identifier 0x81\n drm/nouveau: ensure we\u0027ve parsed i2c table entry for INIT_*I2C* handlers\n drm/nouveau: display error message for any failed init table opcode\n drm/nouveau: fix init table handlers to return proper error codes\n drm/nv50: support fractional feedback divider on newer chips\n drm/nv50: fix monitor detection on certain chipsets\n drm/nv50: store full dcb i2c entry from vbios\n drm/nv50: fix suspend/resume with DP outputs\n drm/nv50: output calculated crtc pll when debugging on\n drm/nouveau: dump pll limits entries when debugging is on\n drm/nouveau: bios parser fixes for eDP boards\n drm/nouveau: fix a nouveau_bo dereference after it\u0027s been destroyed\n drm/nv40: remove some completed ctxprog TODOs\n ...\n" }, { "commit": "7c5ee5366f79f53de2a11e73953daee6d58df124", "tree": "b687002e761917b91e17fed3f327015939b115c8", "parents": [ "8f5b5e632cd55d9acf10ba498b858fd996bd1a39" ], "author": { "name": "Matthew Garrett", "email": "mjg@redhat.com", "time": "Mon Apr 26 16:00:09 2010 -0400" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue May 18 18:21:20 2010 +1000" }, "message": "ttm: Provide an API for starting and stopping the delayed workqueue\n\nWe want to be able to prevent the delayed workqueue from changing state\nwhile we\u0027re reclocking, so add an API to block and unblock it.\n\nSigned-off-by: Matthew Garrett \u003cmjg@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "8cfe92d683a0041ac8e016a0b0a487c99a78f6c1", "tree": "5949c336ea881c9fb51aee811432125451fef45d", "parents": [ "5be6eff965aee22181d7f6210c2c463420e94e03" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Wed Apr 28 11:33:25 2010 +0200" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri May 07 09:21:28 2010 +1000" }, "message": "drm/ttm: Remove the ttm_bo_block_reservation() function.\n\nIt\u0027s unused and buggy in its current form, since it can place a bo\nin the reserved state without removing it from lru lists.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "7547a917fa5f3b2406f52c7dcf7ec9ad3c8532eb", "tree": "59b0d0e9b6c251c4df5799b93395454592004d57", "parents": [ "a8089e849a32c5b6bfd6c88dbd09c0ea4a779b71", "6b8b1786a8c29ce6e32298b93ac8d4a18a2b11c4" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 20 14:15:09 2010 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 20 14:15:09 2010 +1000" }, "message": "Merge branch \u0027drm-ttm-unmappable\u0027 into drm-core-next\n\n* drm-ttm-unmappable:\n drm/radeon/kms: enable use of unmappable VRAM V2\n drm/ttm: remove io_ field from TTM V6\n drm/vmwgfx: add support for new TTM fault callback V5\n drm/nouveau/kms: add support for new TTM fault callback V5\n drm/radeon/kms: add support for new fault callback V7\n drm/ttm: ttm_fault callback to allow driver to handle bo placement V6\n drm/ttm: split no_wait argument in 2 GPU or reserve wait\n\nConflicts:\n\tdrivers/gpu/drm/nouveau/nouveau_bo.c\n" }, { "commit": "0c321c79627189204d7d0bf65ab19f5ac419abed", "tree": "9df52ae46164bdca3dab6a36d5ec17360c60b9e4", "parents": [ "96bf8b8778976a6e6a4fe4e6e0421d8ed7892798" ], "author": { "name": "Jerome Glisse", "email": "jglisse@redhat.com", "time": "Wed Apr 07 10:21:27 2010 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 20 14:13:09 2010 +1000" }, "message": "drm/ttm: remove io_ field from TTM V6\n\nAll TTM driver have been converted to new io_mem_reserve/free\ninterface which allow driver to choose and return proper io\nbase, offset to core TTM for ioremapping if necessary. This\npatch remove what is now deadcode.\n\nV2 adapt to match with change in first patch of the patchset\nV3 update after io_mem_reserve/io_mem_free callback balancing\nV4 adjust to minor cleanup\nV5 remove the needs ioremap flag\nV6 keep the ioremapping facility in TTM\n\n[airlied- squashed driver removals in here also]\n\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nReviewed-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "82c5da6bf8b55a931b042fb531083863d26c8020", "tree": "88168d32d7060598ac730c30967b2e62dc5da28d", "parents": [ "9d87fa2138d06ff400551800d67d522625033e35" ], "author": { "name": "Jerome Glisse", "email": "jglisse@redhat.com", "time": "Fri Apr 09 14:39:23 2010 +0200" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 20 14:12:05 2010 +1000" }, "message": "drm/ttm: ttm_fault callback to allow driver to handle bo placement V6\n\nOn fault the driver is given the opportunity to perform any operation\nit sees fit in order to place the buffer into a CPU visible area of\nmemory. This patch doesn\u0027t break TTM users, nouveau, vmwgfx and radeon\nshould keep working properly. Future patch will take advantage of this\ninfrastructure and remove the old path from TTM once driver are\nconverted.\n\nV2 return VM_FAULT_NOPAGE if callback return -EBUSY or -ERESTARTSYS\nV3 balance io_mem_reserve and io_mem_free call, fault_reserve_notify\n is responsible to perform any necessary task for mapping to succeed\nV4 minor cleanup, atomic_t -\u003e bool as member is protected by reserve\n mecanism from concurent access\nV5 the callback is now responsible for iomapping the bo and providing\n a virtual address this simplify TTM and will allow to get rid of\n TTM_MEMTYPE_FLAG_NEEDS_IOREMAP\nV6 use the bus addr data to decide to ioremap or this isn\u0027t needed\n but we don\u0027t necesarily need to ioremap in the callback but still\n allow driver to use static mapping\n\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nReviewed-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "9d87fa2138d06ff400551800d67d522625033e35", "tree": "284cd0f73ccb2f2fad1c71f974d4e9e4d0035e81", "parents": [ "3a89b4a9ca7ce11e3b7d5119aea917b9fc29a302" ], "author": { "name": "Jerome Glisse", "email": "jglisse@redhat.com", "time": "Wed Apr 07 10:21:19 2010 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Apr 08 10:21:19 2010 +1000" }, "message": "drm/ttm: split no_wait argument in 2 GPU or reserve wait\n\nThere is case where we want to be able to wait only for the\nGPU while not waiting for other buffer to be unreserved. This\npatch split the no_wait argument all the way down in the whole\nttm path so that upper level can decide on what to wait on or\nnot.\n\n[airlied: squashed these 4 for bisectability reasons.]\ndrm/radeon/kms: update to TTM no_wait splitted argument\ndrm/nouveau: update to TTM no_wait splitted argument\ndrm/vmwgfx: update to TTM no_wait splitted argument\n[vmwgfx patch: Reviewed-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e]\n\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nAcked-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "c96af79e3463d5d3f865625baa8bb8aa4c0944a0", "tree": "d3ed3dd9a4a9e8beda500f51213bd0c7e5c28884", "parents": [ "975efdb1bf925ad48d4e3fe5339a85f12601e10d" ], "author": { "name": "Pauli Nieminen", "email": "suokkos@gmail.com", "time": "Thu Apr 01 12:45:03 2010 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 06 11:36:18 2010 +1000" }, "message": "drm/ttm: Add sysfs interface to control pool allocator.\n\nSysfs interface allows user to configure pool allocator functionality and\nchange limits for the size of pool.\n\nSigned-off-by: Pauli Nieminen \u003csuokkos@gmail.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "0745866165598b067442c472911280527b08be3e", "tree": "454900be36c969fe45554a47e733807d63be8b6f", "parents": [ "1403b1a38e8b19a4cc17e2c158e278628943a436" ], "author": { "name": "Pauli Nieminen", "email": "suokkos@gmail.com", "time": "Thu Apr 01 12:44:58 2010 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 06 11:35:45 2010 +1000" }, "message": "drm/ttm: Add debugfs output entry to pool allocator.\n\nttm_page_alloc_debugfs can be registered to output the state\nof pools.\n\nDebugfs file will output number of pages freed from the pool,\nnumber of pages in pool now and the lowes number of pages in\npool since previous shrink.\n\nSigned-off-by: Pauli Nieminen \u003csuokkos@gmail.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "1403b1a38e8b19a4cc17e2c158e278628943a436", "tree": "309781b3948092d97e47e64eab9f14db85c357cd", "parents": [ "95beb690170e6ce918fe53c73a0fcc7cf64d704a" ], "author": { "name": "Pauli Nieminen", "email": "suokkos@gmail.com", "time": "Thu Apr 01 12:44:57 2010 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Apr 06 11:35:26 2010 +1000" }, "message": "drm/ttm: add pool wc/uc page allocator V3\n\nOn AGP system we might allocate/free routinely uncached or wc memory,\nchanging page from cached (wb) to uc or wc is very expensive and involves\na lot of flushing. To improve performance this allocator use a pool\nof uc,wc pages.\n\nPools are protected with spinlocks to allow multiple threads to allocate pages\nsimultanously. Expensive operations are done outside of spinlock to maximize\nconcurrency.\n\nPools are linked lists of pages that were recently freed. mm shrink callback\nallows kernel to claim back pages when they are required for something else.\n\nFixes:\n* set_pages_array_wb handles highmem pages so we don\u0027t have to remove them\n from pool.\n* Add count parameter to ttm_put_pages to avoid looping in free code.\n* Change looping from _safe to normal in pool fill error path.\n* Initialize sum variable and make the loop prettier in get_num_unused_pages.\n\n* Moved pages_freed reseting inside the loop in ttm_page_pool_free.\n* Add warning comment about spinlock context in ttm_page_pool_free.\n\nBased on Jerome Glisse\u0027s and Dave Airlie\u0027s pool allocator.\n\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\nSigned-off-by: Pauli Nieminen \u003csuokkos@gmail.com\u003e\nReviewed-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "72e942dd846f98e2d35aad5436d77a878ef05c5e", "tree": "bed31b516da9fd80058800caa6161b6e0f880595", "parents": [ "d424b925f7092b9d95e0a8556872349abe79d9b6" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Mar 09 06:33:26 2010 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Mar 15 10:31:43 2010 +1000" }, "message": "drm/ttm: use drm calloc large and free large\n\nNow that the drm core can do this, lets just use it, split the code out\nso TTM doesn\u0027t have to drag all of drmP.h in.\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "a55e8d452ed2f6bbecda1a3039e82cd05244be3d", "tree": "4865b9575e3d6c22b68b2df70eaec27c37f73b36", "parents": [ "d964fc54ed06cea45dfd10832ed3d34f3ddb661b" ], "author": { "name": "Randy Dunlap", "email": "randy.dunlap@oracle.com", "time": "Wed Feb 24 14:29:14 2010 -0800" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Mar 01 16:03:26 2010 +1000" }, "message": "drm/ttm: fix function prototype to match implementation\n\nFix function prototype to match its actual usage and implementation.\n\ndrivers/gpu/drm/ttm/ttm_bo_util.c:341:10: error: symbol \u0027ttm_io_prot\u0027 redeclared with different type (originally declared at include/drm/ttm/ttm_bo_driver.h:911) - incompatible argument 1 (different signedness)\n\nSigned-off-by: Randy Dunlap \u003crandy.dunlap@oracle.com\u003e\nCc:\tDavid Airlie \u003cairlied@linux.ie\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "3f09ea4ecdcbcea05541f83e557d6ce2e56626d8", "tree": "445e4519b2e8a39136f611f07db670d1e10ace6e", "parents": [ "effe1105bef07f42366c20eac41b80ff9fcf675e" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Wed Jan 13 22:28:40 2010 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Jan 14 12:18:54 2010 +1000" }, "message": "drm/ttm: Add a swap_notify callback.\n\nThis is needed for a bugfix in the vmwgfx driver.\nDrivers may have GPU bindings on buffers that core TTM is not aware of,\nand TTM may view those buffers as ordinary system memory buffers.\nAdd a notifier to such drivers when TTM is about to move the buffer\ncontents out to swappable memory. The driver must then release any\nprivate GPU bindings on those buffers.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "be1cb8689c480228ffd2e4bfccc0dab7156cd9ea", "tree": "b2354c0b21e89f627a5bc8f9cf5ba5ad0bdcb881", "parents": [ "d1ede145cea25c5b6d2ebb19b167af14e374bb45" ], "author": { "name": "Jakob Bornecrantz", "email": "jakob@vmware.com", "time": "Mon Dec 14 22:07:45 2009 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Tue Dec 15 08:38:18 2009 +1000" }, "message": "drm/ttm: Add more driver type enums\n\nSigned-off-by: Jakob Bornecrantz \u003cjakob@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "09855acb1c2e3779f25317ec9a8ffe1b1784a4a8", "tree": "4a96a67c6bc9694036a95aafd4004c9e9b89eadc", "parents": [ "4361e52ad0372e6fd2240a2207b49a4de1f45ca9" ], "author": { "name": "Jerome Glisse", "email": "jglisse@redhat.com", "time": "Thu Dec 10 17:16:27 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Dec 11 14:09:05 2009 +1000" }, "message": "drm/ttm: Convert ttm_buffer_object_init to use ttm_placement\n\nConvert ttm_buffer_object_init to use struct ttm_placement and\nrename to ttm_bo_init for consistency with function naming. This\nallow to give more complex placement at buffer creation. For\ninstance you ask to allocate bo into vram first but if there is\nnot enough vram you can give system as a second possible\nplacement. It also allow to create buffer in a specific range.\n\nAlso rename ttm_buffer_object_validate to ttm_bo_validate.\n\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "98ffc4158e12008102cb6ae242a7fc46f9243f0d", "tree": "fa8b231281fac87afe88ab9a0539c6ab5cdc2100", "parents": [ "312ea8da049a1830aa50c6e00002e50e30df476e" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Mon Dec 07 18:36:18 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Dec 10 15:09:03 2009 +1000" }, "message": "drm/ttm: Have the TTM code return -ERESTARTSYS instead of -ERESTART.\n\nReturn -ERESTARTSYS instead of -ERESTART when interrupted by a signal.\nThe -ERESTARTSYS is converted to an -EINTR by the kernel signal layer\nbefore returned to user-space.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "ca262a9998d46196750bb19a9dc4bd465b170ff7", "tree": "b128691e5c57f6305c5752ac5c1b09e6aedfb650", "parents": [ "a2e68e92d384d37c8cc6bb7206d43b1eb9bc3f08" ], "author": { "name": "Jerome Glisse", "email": "jglisse@redhat.com", "time": "Tue Dec 08 15:33:32 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Dec 10 15:09:02 2009 +1000" }, "message": "drm/ttm: Rework validation \u0026 memory space allocation (V3)\n\nThis change allow driver to pass sorted memory placement,\nfrom most prefered placement to least prefered placement.\nIn order to avoid long function prototype a structure is\nused to gather memory placement informations such as range\nrestriction (if you need a buffer to be in given range).\nRange restriction is determined by fpfn \u0026 lpfn which are\nthe first page and last page number btw which allocation\ncan happen. If those fields are set to 0 ttm will assume\nbuffer can be put anywhere in the address space (thus it\navoids putting a burden on the driver to always properly\nset those fields).\n\nThis patch also factor few functions like evicting first\nentry of lru list or getting a memory space. This avoid\ncode duplication.\n\nV2: Change API to use placement flags and array instead\n of packing placement order into a quadword.\nV3: Make sure we set the appropriate mem.placement flag\n when validating or allocation memory space.\n\n[Pending Thomas Hellstrom further review but okay\nfrom preliminary review so far].\n\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "4bfd75cb08a362cb1df35dc6a5032d12843c6d87", "tree": "1363edb589855b593d916411bf3d2f2b5e6cfcf1", "parents": [ "c078aa2fc4d8e022c3b611e07b25ff77afdf9b73" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Sun Dec 06 21:46:27 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Dec 07 15:22:07 2009 +1000" }, "message": "drm/ttm: Export symbols needed for the vmwgfx driver.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "c078aa2fc4d8e022c3b611e07b25ff77afdf9b73", "tree": "60349ecc7344001b3824529be376b4677012583d", "parents": [ "4aff1013f5e4ae08a24155c029a2c5e1a7929de6" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Sun Dec 06 21:46:26 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Dec 07 15:22:05 2009 +1000" }, "message": "drm/ttm: Add TTM execbuf utilities.\n\nUtilities to reserve, unreserve and fence a list of TTM\nbuffer objects in a deadlock-safe manner.\n\nUsed by the vmwgfx driver.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "4aff1013f5e4ae08a24155c029a2c5e1a7929de6", "tree": "1317322a9e6e4c6094cfa32090c05c20db252ec0", "parents": [ "88071539a3f5195f9e9dae38a3e35b3ce4b9f9fc" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Sun Dec 06 21:46:25 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Dec 07 15:22:03 2009 +1000" }, "message": "drm/ttm: Add ttm lock functionality.\n\nThis is intended to be used by ttm-aware drivers to\n1) Block clients to inactive masters when\nthey try to validate buffers for GPU use.\n2) Optionally block clients to the current master when\nthere is thrashing due to GPU memory shortage.\n\nUsed by the vmwgfx driver.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "88071539a3f5195f9e9dae38a3e35b3ce4b9f9fc", "tree": "a0be303c46b4cfa28b8357cf27ab6fae666fe470", "parents": [ "01d01ba947670cf58f22119fc126fdf39078f6ba" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Sun Dec 06 21:46:24 2009 +0100" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Dec 07 15:22:01 2009 +1000" }, "message": "drm/ttm: Add user-space objects.\n\nAdd objects needed for user-space to maintain reference counts on ttm objects.\nThis is used by the vmwgfx driver which allows user-space to maintain\nmap-counts on dma buffers, lock-counts on the ttm lock and ref-counts on\ngpu surfaces, gpu contexts and dma buffer.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "c3a73ba13bac7fd96030f39202b2d37fb19c46a6", "tree": "e91c424da8f71a5e5923ff4175c3b15cf4543b65", "parents": [ "46a79fa08a9a890a12cf9ec3ce51800911a907bf" ], "author": { "name": "Martin Michlmayr", "email": "tbm@cyrius.com", "time": "Thu Nov 19 16:29:45 2009 +0000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Fri Dec 04 09:11:16 2009 +1000" }, "message": "drm/ttm: Fix build failure due to missing struct page\n\ndrm/ttm fails to build on MIPS because \"struct page\" is not known:\n| In file included from drivers/gpu/drm/ttm/ttm_memory.c:28:\n| include/drm/ttm/ttm_memory.h:154: warning: \u0027struct page\u0027 declared inside parameter list\n| include/drm/ttm/ttm_memory.h:154: warning: its scope is only this definition or declaration, which is probably not what you want\n| include/drm/ttm/ttm_memory.h:156: warning: \u0027struct page\u0027 declared inside parameter list\n| drivers/gpu/drm/ttm/ttm_memory.c:540: error: conflicting types for \u0027ttm_mem_global_alloc_page\u0027\n| include/drm/ttm/ttm_memory.h:154: error: previous declaration of \u0027ttm_mem_global_alloc_page\u0027 was here\n| drivers/gpu/drm/ttm/ttm_memory.c:561: error: conflicting types for \u0027ttm_mem_global_free_page\u0027\n| include/drm/ttm/ttm_memory.h:156: error: previous declaration of \u0027ttm_mem_global_free_page\u0027 was here\n\nSigned-off-by: Martin Michlmayr \u003ctbm@cyrius.com\u003e\nCc: stable@kernel.org\nAcked-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "51c8b4071d84d46cc100baa5931ad06b2a823c95", "tree": "098cf9d41ce1c548d922708a770a9efe35e434df", "parents": [ "a987fcaa805fcb24ba885c2e29fd4fdb6816f08f", "6c30c53fd5ae6a99a23ad78e90c428d2c8ffb07f" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Aug 20 13:38:04 2009 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Thu Aug 20 13:38:04 2009 +1000" }, "message": "Merge Linus master to drm-next\n\nlinux-next conflict reported needed resolution.\n\nConflicts:\n\tdrivers/gpu/drm/drm_crtc.c\n\tdrivers/gpu/drm/drm_edid.c\n\tdrivers/gpu/drm/i915/intel_sdvo.c\n\tdrivers/gpu/drm/radeon/radeon_ttm.c\n\tdrivers/gpu/drm/ttm/ttm_bo.c\n" }, { "commit": "a987fcaa805fcb24ba885c2e29fd4fdb6816f08f", "tree": "561b6dd8e002e2eb1a75132b1edbd303782dc2fb", "parents": [ "5fd9cbad3a4ae82c83c55b9c621d156c326724ef" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Tue Aug 18 16:51:56 2009 +0200" }, "committer": { "name": "Dave Airlie", "email": "airlied@linux.ie", "time": "Wed Aug 19 16:10:34 2009 +1000" }, "message": "ttm: Make parts of a struct ttm_bo_device global.\n\nCommon resources, like memory accounting and swap lists should be\nglobal and not per device. Introduce a struct ttm_bo_global to\naccomodate this, and register it with sysfs. Add a small sysfs interface\nto return the number of active buffer objects.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@linux.ie\u003e\n" }, { "commit": "5fd9cbad3a4ae82c83c55b9c621d156c326724ef", "tree": "1a0868a3bd2751fa861c083aeb3ac27f3f695694", "parents": [ "e9840be8c23601285a70520b4898818f28ce8c2b" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Mon Aug 17 16:28:39 2009 +0200" }, "committer": { "name": "Dave Airlie", "email": "airlied@linux.ie", "time": "Wed Aug 19 16:09:53 2009 +1000" }, "message": "drm/ttm: Memory accounting rework.\n\nUse inclusive zones to simplify accounting and its sysfs representation.\nUse DMA32 accounting where applicable.\n\nAdd a sysfs interface to make the heuristically determined limits\nreadable and configurable.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@linux.ie\u003e\n" }, { "commit": "a0724fcf829e5afb66159ef68cb16a805ea11b42", "tree": "f407795b0b6e38b8cbdf141de4135810c6a5ceab", "parents": [ "949ef70e2d1a5c12178875f513df34fc85d91a38" ], "author": { "name": "Pekka Paalanen", "email": "pq@iki.fi", "time": "Mon Aug 17 01:18:38 2009 +0300" }, "committer": { "name": "Dave Airlie", "email": "airlied@linux.ie", "time": "Wed Aug 19 16:07:00 2009 +1000" }, "message": "drm/ttm: optimize bo_kmap_type values\n\nA micro-optimization on the function ttm_kmap_obj_virtual().\n\nBy defining the values of enum ttm_bo_kmap_obj::bo_kmap_type to have a\nbit indicating iomem, size of the function ttm_kmap_obj_virtual() will be\nreduced by 16 bytes on x86_64 (gcc 4.1.2).\n\nttm_kmap_obj_virtual() may be heavily used, when buffer objects are\naccessed via wrappers, that work for both kinds of memory addresses:\niomem cookies and kernel virtual.\n\nSigned-off-by: Pekka Paalanen \u003cpq@iki.fi\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "e024e11070a0a0dc7163ce1ec2da354a638bdbed", "tree": "add483e7428f91da6f3c26be702aa45e6d69b694", "parents": [ "c836e862803b2aa2bd9a354e151316d2b42c44ec" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jun 24 09:48:08 2009 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jul 29 15:42:18 2009 +1000" }, "message": "drm/radeon/kms: add initial colortiling support.\n\nThis adds new set/get tiling interfaces where the pitch\nand macro/micro tiling enables can be set. Along with\na flag to decide if this object should have a surface when mapped.\n\nThe only thing we need to allocate with a mapped surface should be\nthe frontbuffer. Note rotate scanout shouldn\u0027t require one, and\nback/depth shouldn\u0027t either, though mesa needs some fixes.\n\nIt fixes the TTM interfaces along Thomas\u0027s suggestions, and I\u0027ve tested\nthe surface stealing code with two X servers and not seen any lockdep issues.\n\nI\u0027ve stopped tiling the fbcon frontbuffer, as I don\u0027t see there being\nany advantage other than testing, I\u0027ve left the testing commands in there,\njust flip the fb_tiled to true in radeon_fb.c\n\nOpen: Can we integrate endian swapping in with this?\n\nFuture features:\ntexture tiling - need to relocate texture registers TXOFFSET* with tiling info.\n\nThis also merges Michel\u0027s cleanup surfaces regs at init time patch\neven though it makes sense on its own, this patch really relies on it.\n\nSome PowerMac firmwares set up a tiling surface at the beginning of VRAM\nwhich messes us up otherwise.\nthat patch is:\nSigned-off-by: Michel Dänzer \u003cdaenzer@vmware.com\u003e\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "ad49f501867cba87e1e45e5ebae0b12435d68bf1", "tree": "4602e5caf96451b1dcdda7a38628d494466d2e20", "parents": [ "61b576dbbe6a19d102c025ebc102a0749e2d3c80" ], "author": { "name": "Dave Airlie", "email": "airlied@linux.ie", "time": "Fri Jul 10 22:36:26 2009 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jul 15 17:13:18 2009 +1000" }, "message": "drm/ttm/radeon: add dma32 support.\n\nThis add support for using dma32 memory on gpus that really need it.\n\nCurrently IGPs are left without DMA32 but we might need to change\nthat unless we can fix rs690.\n\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "d1724078d6a01177c1db4ea0b75fda1ca8a73d57", "tree": "24c2cef55a572938acadbbd662df4d65701ca38e", "parents": [ "531369e62649bb8f31217cc0bf33ee6f89f1dff6" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Wed Jun 24 19:57:35 2009 +0200" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jul 15 17:13:09 2009 +1000" }, "message": "ttm: Make messages more readable.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" }, { "commit": "ba4e7d973dd09b66912ac4c0856add8b0703a997", "tree": "32a87edb83a427ffd22645c5f77e6cec8be4e719", "parents": [ "e6c03c5b40314d787f7053f631594d6b1bd609e8" ], "author": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Wed Jun 10 15:20:19 2009 +0200" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Mon Jun 15 09:37:57 2009 +1000" }, "message": "drm: Add the TTM GPU memory manager subsystem.\n\nTTM is a GPU memory manager subsystem designed for use with GPU\ndevices with various memory types (On-card VRAM, AGP,\nPCI apertures etc.). It\u0027s essentially a helper library that assists\nthe DRM driver in creating and managing persistent buffer objects.\n\nTTM manages placement of data and CPU map setup and teardown on\ndata movement. It can also optionally manage synchronization of\ndata on a per-buffer-object level.\n\nTTM takes care to provide an always valid virtual user-space address\nto a buffer object which makes user-space sub-allocation of\nbig buffer objects feasible.\n\nTTM uses a fine-grained per buffer-object locking scheme, taking\ncare to release all relevant locks when waiting for the GPU.\nAlthough this implies some locking overhead, it\u0027s probably a big\nwin for devices with multiple command submission mechanisms, since\nthe lock contention will be minimal.\n\nTTM can be used with whatever user-space interface the driver\nchooses, including GEM. It\u0027s used by the upcoming Radeon KMS DRM driver\nand is also the GPU memory management core of various new experimental\nDRM drivers.\n\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\nSigned-off-by: Jerome Glisse \u003cjglisse@redhat.com\u003e\nSigned-off-by: Dave Airlie \u003cairlied@redhat.com\u003e\n" } ] }