)]}' { "log": [ { "commit": "8c643ac79be63aba114fd7682297858354c198b9", "tree": "d88092a27ddca4f4905c301d4e91c63dc97f883d", "parents": [ "f89945b6cdf14caccf607b6ee6197402945c0d90" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jul 17 10:33:10 2020 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jul 17 10:41:28 2020 -0700" }, "message": "gldownloadelement: fix export of BGR/BGRA textures\n\nGStreamer always stores data in RGBA textures, then uses shaders to\nswizzle and convert between formats. When we get a texture GStreamer\nsays is BGR, it\u0027s really an RGB texture as far as the driver is\nconcerned. If we ask the driver to export it as BGR, driver will\nsee that source is RGB (according to its internal bookkeeping) and\ndestination is BGR, so it swizzles B and R. When GStreamer gets the\noutput intended to be BGR, it\u0027s therefore really RGB because it\nwas swizzled when texture was generated, and then again on export.\n\nSolution is simple; always export the data as RGB/A to get the bytes\nexactly as GStreamer stored them in the texture, as they\u0027re already in\nthe correct format.\n\nBug: 161191133\nChange-Id: Ia63a5c585d4e70baef481929f0a71f96d0a80701\n" }, { "commit": "f89945b6cdf14caccf607b6ee6197402945c0d90", "tree": "5e9fe985c289525bea8566fe5385deb8f0159711", "parents": [ "61dddf47f6bf52ee3a80eec6c249fc83ac86c35b" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jul 15 16:36:51 2020 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jul 15 16:36:51 2020 -0700" }, "message": "eglimage: allow direct import of RGB dma-bufs\n\nAdd mapping between GStreamer and DRM format for direct import\nof RGB pixel data.\n\nBug: 161191133\nChange-Id: I6b2438f457bb38107b10e4bdfe175d83b4025800\n" }, { "commit": "61dddf47f6bf52ee3a80eec6c249fc83ac86c35b", "tree": "50989fdaae9997be716c05417f30addf05e6663b", "parents": [ "1c6816d5bb0ee5618f093623dabc0f1bdd985ede" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jul 15 16:31:29 2020 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jul 15 16:31:29 2020 -0700" }, "message": "glcolorconvert: fix non-passthrough of RGB -\u003e RGB\n\nOn i.MX we\u0027re always drawing EGLImages once since textures\nare imported from dma-bufs that can\u0027t be re-exported.\nWhen input and output are both RGB GStreamer tries to compile\na shader for the purpose, but the alpha channel mapping is\nmissing when\n\n- input is same format as output\n- that same format is only 3 chars\n\nThe channel map must always be 4 chars long, one char for\neach of R G B A.\n\nBug: 161191133\nChange-Id: Ib8d19fb30f09ef8ff24260014a05bb24e0a13b44\n" }, { "commit": "1c6816d5bb0ee5618f093623dabc0f1bdd985ede", "tree": "b453f6efc3ffe29016ff79eae1e61fdbb13e3089", "parents": [ "48cafd291b975fd5012b19845cc7b2a5830ffe89" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Jun 02 10:56:54 2020 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jun 03 12:05:44 2020 -0700" }, "message": "gldownloadelement: fix export of odd width textures\n\nNew GPU driver changed the default stride calculations when resolving\ntextures to external surfaces. To get tightly packed RGB out of i.MX8MQ we\nnow must specify the stride value explicitly.\n\nBUG: 157921487\nChange-Id: I0ef6bf1cfc3230ff8a4097e6db0766f04a100f53\n" }, { "commit": "48cafd291b975fd5012b19845cc7b2a5830ffe89", "tree": "c659ba2472ca8f02ce0a858b1c65cd6b6d37cc15", "parents": [ "8ab9b3ee5ae974dccbf30d85b6d3c4bfe9ff4709" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jan 22 11:59:53 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jan 24 16:52:28 2020 -0800" }, "message": "gldownloadelement: fix downloading of \u0027RGBA\u0027 EGLImages\n\nNXP has broken basic standard use like\nsource ! glupload ! glcolorconvert ! gldownload ! video/x-raw,format\u003dRGBA ! sink\nin their fork. In this pipeline, if source produces dmabufs, the\npipeline output is always just black. That\u0027s because the dmabuf wrapped\nEGLImage marked as RGBA is never drawn. glcolorconvert will pass through\nof same RGBA caps but the buffer isn\u0027t RGBA in traditional GStreamer\nsense, it can simply just be sampled as a RGBA texture by hardware when\ndrawing.\n\nDetect this situation and force drawing to an actual RGBA texture that\ncan be efficiently downloaded. This restores standard GStreamer\nbehavior.\n\nBUG: 148157331\n\nChange-Id: I229efde22f49d9540c85b38923dbea4d4eb257bc\n" }, { "commit": "8ab9b3ee5ae974dccbf30d85b6d3c4bfe9ff4709", "tree": "4e9001328898c2f6d17d9399095c870eced54240", "parents": [ "e390164e76ea169b0ee1adb1581f55316422510e" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jan 17 14:51:55 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jan 24 16:44:18 2020 -0800" }, "message": "gldownloadelement: accelerate with glExportTexImageDMA\n\nAccelerate GL texture download with glExportTexImageDMA if driver\nsupports it.\n\nOn columbia, dowloading 720p textures:\n* RGBA performace up from 28 fps to 108 fps (3.8x improvement)\n* RGB performance up from 2 fps to 48 fps (24x improvement)\n\nEnterprise sees performance improvments too but they\u0027re less drastic\nas it\u0027s a much powerful platform overall.\nMeasurements based on otherwise idle system, add more workloads such\nas on screen rendering and performance increase will not be as\napparent.\n\nBUG: 146907014\nChange-Id: Idced5ea3aeae835187461bef0cf0d9113f27bfe3\n" }, { "commit": "e390164e76ea169b0ee1adb1581f55316422510e", "tree": "6f74fb38c5e2d79b323b29d32e030e61ca1eeb72", "parents": [ "a9879f9339e838a99693b18a2bd0df1b9555f585" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Jan 23 14:21:10 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Sat Jan 25 00:23:44 2020 +0000" }, "message": "dmabuf: always map dmabuf with PROT_WRITE\n\nThis is to enable zero copy inference on dmabufs. Gasket kernel driver\nrequires all user pointers passed to it to be writable in\ndrivers/staging/gasket/gasket_page_table.c gasket_perform_mapping\neven if it\u0027s not actually writing to the buffer. If we map a buffer\nwith GST_MAP_READ it\u0027ll be mmapped with PROT_READ and gasket will\nerror out with -EFAULT.\n\nWhile a more correct solution is to fix the gasket driver the needed\nlinux/mm APIs aren\u0027t there in the current version of the kernel.\nWe\u0027d need\nhttps://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id\u003d73b0140bf0fe9df90fb267c00673c4b9bf285430\nwhich is for a much later kernel version.\n\nThis change may look dangerous but what it really does is to make\ndmabuf protection equal to buffers allocated by the system memory\nallocator; i.e. even if you map a buffer with GST_MAP_READ only\nyou can, but should not, write to it.\n\nBUG: 142164990\nBUG: 148226798\nBUG: 148221549\nChange-Id: Ie7f7cbce212fd6851e9047e05245b9f634c5e5fd\n" }, { "commit": "a9879f9339e838a99693b18a2bd0df1b9555f585", "tree": "ef6f05a2cc93dfac11bc660efe60a8621bf2f62d", "parents": [ "af8dfd410589efbbe7ea60f9b69ca45f72738db7" ], "author": { "name": "Nicolas Dufresne", "email": "nicolas.dufresne@collabora.com", "time": "Wed Mar 21 18:15:49 2018 -0400" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Sat Jan 25 00:23:30 2020 +0000" }, "message": "dmabufallocator: Fix build if LINUX_DMA_BUF_H is missing\n\nBUG: 148221549\nChange-Id: I620f68b0884ef2da341fe4f651a04090ec67e83b\n" }, { "commit": "af8dfd410589efbbe7ea60f9b69ca45f72738db7", "tree": "d47c7106e28ada8edb6a4dee022392816005e626", "parents": [ "610169870dea636c690184a3b59b444c63dd8042" ], "author": { "name": "Nicolas Dufresne", "email": "nicolas.dufresne@collabora.com", "time": "Fri Mar 09 21:24:52 2018 -0500" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Sat Jan 25 00:23:04 2020 +0000" }, "message": "allocators: Add DMABuf synchronization\n\nWhen doing CPU Access, some architecture may require caches to be\nsynchronize before use. Otherwise, some visual artifact may be\nvisible, as the CPU modification may still resides in cache.\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d794216\n\nBUG: 148221549\nChange-Id: I0d54d8368d9ed5855721852493183ee88205929d\n" }, { "commit": "610169870dea636c690184a3b59b444c63dd8042", "tree": "20c928f7fc177027186a7106a71c90ea35f54661", "parents": [ "3221a095ca45d957a54791cd58dce014fb67ff6c" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Jan 23 10:54:39 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Sat Jan 25 00:22:30 2020 +0000" }, "message": "Revert \"fdmem: always map dmabuf with PROT_WRITE\"\n\nWill be replaced with a similar change.\n\nBUG: 148221549\n\nThis reverts commit 221717a7f60b4793007077e3aa7fe29bc1dc11eb.\n\nChange-Id: I52f3ecaafcd58fe95318f6229f1c81508fdd0c8c\n" }, { "commit": "3221a095ca45d957a54791cd58dce014fb67ff6c", "tree": "200483f34611fbed6ca5ced16369f77c2a274663", "parents": [ "b4d4947f758c84ee2ba658d5c0a94cd116814d2b" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jan 17 14:48:00 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Jan 23 14:25:36 2020 -0800" }, "message": "video-converter: reduce minimum lines per CPU core\n\nFor multi CPU core video conversion there\u0027s a explicitly magic minimum\nnumber of lines to convert per core, set at 200 by upstream. This is\nreally a tunable that depends on CPU horsepower and use case.\n\nOn our systems resolutions converted on CPU is generally low, typically\n200 - 300 lines high. That means that CPU conversion only runs on 1 - 2\ncores even if usage of all cores were requested. Reduce minimum lines\nper core to 50 so that all cores are utilized converting a 200 lines\nhigh frame on a quad core system.\n\nChange-Id: I56894c64271437c637c5541e489f779ecc654559\n" }, { "commit": "b4d4947f758c84ee2ba658d5c0a94cd116814d2b", "tree": "e32b86fca7fa4946e4b2764544d27a8944fbe51d", "parents": [ "d40df83b5e513a05d4a79e2e766735f9f86c9dda" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jan 17 14:41:46 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Jan 17 14:41:46 2020 -0800" }, "message": "glmemorydma: fix build warnings in NXP code\n\nChange-Id: I29926ba8228b1aea9dafe3d7910606c246acbcac\n" }, { "commit": "d40df83b5e513a05d4a79e2e766735f9f86c9dda", "tree": "b845149a1ca910248af5c9b666a16a770dfee1ba", "parents": [ "f6f3a0cf3a8d277295243e21a8d3204f0c428809" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jan 15 13:42:42 2020 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Jan 15 13:42:42 2020 -0800" }, "message": "gldownloadelement: fix memory leak and build warnings\n\n* gst_pad_get_current_caps ups the refcount on returned caps, so\n caller must unref or the caps leak\n* fix build warnings related to GstMemory by casting correctly\n\nChange-Id: I113aa1bc86fe42f3c588483fea401d9c1421446f\n" }, { "commit": "f6f3a0cf3a8d277295243e21a8d3204f0c428809", "tree": "cde54ea9b8ab5fe718bef6a012db663d6256b3b9", "parents": [ "ccfbf11dc3065e3aa4e5d2d30cd4beeb1b27a429" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Dec 26 16:04:12 2019 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Dec 26 16:04:12 2019 -0800" }, "message": "glmemorydma: wait for completion with gst_gl_sync_meta_wait_cpu\n\nRequired on Columbia. Its blitter is different from Enterprise and\nuses a different code path under the hood. NOP on Enterprise.\n\nChange-Id: I2738a20cc07b996f7f3da55d8f553c1c904583a6\n" }, { "commit": "ccfbf11dc3065e3aa4e5d2d30cd4beeb1b27a429", "tree": "6e5c0481457761b82228559c6c60c21b85a94cf5", "parents": [ "d4fc316ac2771b99c4195675e5636a65c71441b1" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Nov 12 13:53:24 2019 -0800" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Nov 12 13:53:24 2019 -0800" }, "message": "glimagesink: full screen should be disabled by default\n\nThis is to be compatible with the default upstream behavior.\n\nBUG: 142900877\nChange-Id: Id2687bb05af260503cadbffc5159dec48d2191c7\n" }, { "commit": "d4fc316ac2771b99c4195675e5636a65c71441b1", "tree": "2ea61f10667b778f41d50666331bfaf05a034602", "parents": [ "29b529465994db81ca6a1e8eff6c9fe0a819b046" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Oct 18 10:20:15 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Oct 18 10:21:16 2019 -0700" }, "message": "glimagesink: wayland: fix build\n\nBUG: 142900877\n\nChange-Id: I62f17dcf9ff516fc9d8ebdcb32d821a139ba5038\n" }, { "commit": "29b529465994db81ca6a1e8eff6c9fe0a819b046", "tree": "6720b0f074526044a864da5e525bc357ee38d02c", "parents": [ "4340a2c2e4f8af22e9f32453b1bf283310028dca" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Oct 17 16:09:00 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Oct 17 16:09:00 2019 -0700" }, "message": "glimagesink: add property to request full screen window\n\nOnly affects windows created internally, externally created windows\n(i.e. windows created and owned by apps) are not affected.\n\nBUG: 142900877\nChange-Id: I8b3039f0b063f7b46e4040419e750c057f185ced\n" }, { "commit": "4340a2c2e4f8af22e9f32453b1bf283310028dca", "tree": "e9dd87a7969609049fd56890883244a196efaa86", "parents": [ "221717a7f60b4793007077e3aa7fe29bc1dc11eb" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Oct 15 00:50:25 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Oct 15 00:50:25 2019 -0700" }, "message": "Tightly pack RGB buffers downloaded from GPU.\n\nGStreamer defaults to 4 byte aligning stride for RGB/BGR whereas\nstride \u003d\u003d width * bpp is expected by networks. GPU can support this\nso override the default stride value when allocating the DMA buffer\nfor download, and propagate the tightly packed stride downstream.\n\nBUG: 142698697\nChange-Id: I0327f746e547f5ad1805f28efcbd9ea0c6512d9f\n" }, { "commit": "221717a7f60b4793007077e3aa7fe29bc1dc11eb", "tree": "c83aa3de1e423a8793e0672d9cfab3973a9321bd", "parents": [ "a77a0ca4a8c7022cf5ef177e41a9dd5a5c95ee54" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Oct 09 12:33:00 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Oct 09 12:40:53 2019 -0700" }, "message": "fdmem: always map dmabuf with PROT_WRITE\n\nThis is to enable zero copy inference on dmabufs. Gasket kernel driver\nrequires all user pointers passed to it to be writable in\ndrivers/staging/gasket/gasket_page_table.c gasket_perform_mapping\neven if it\u0027s not actually writing to the buffer. If we map a buffer\nwith GST_MAP_READ it\u0027ll be mmapped with PROT_READ and gasket will\nerror out with -EFAULT.\n\nWhile a more correct solution is to fix the gasket driver the needed\nlinux/mm APIs aren\u0027t there in the current version of the kernel.\nWe\u0027d need\nhttps://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id\u003d73b0140bf0fe9df90fb267c00673c4b9bf285430\nwhich is for a much later kernel version.\n\nThis change may look dangerous but what it really does is to make\ndmabuf protection equal to buffers allocated by the system memory\nallocator; i.e. even if you map a buffer with GST_MAP_READ only\nyou can, but should not, write to it.\n\nBUG: 142164990\n\nChange-Id: Ie615f923c44969a3606d8856311ffe853a0dd742\n" }, { "commit": "a77a0ca4a8c7022cf5ef177e41a9dd5a5c95ee54", "tree": "3aae31a915a0c55096dfafb5dbf262f5fcc07b53", "parents": [ "5c92dea09585baf385111dad4e4e24e535a2bf0d" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Oct 07 12:50:55 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Oct 07 12:52:10 2019 -0700" }, "message": "Allocate cached ION buffers.\n\nWe\u0027re doing lots of CPU access of ION buffers and with no caching this\naccess is pretty slow. By allocating with ION_FLAG_CACHED CPU mapped\npages will be cached and for example a memcpy of a decoded 1080p video\nframe is about 3x faster. When writing to a buffer from CPU (\"software\nrendering\") ION will ensure proper cache management on munmap.\n\nFor cases not involving CPU access (e.g. VPU directly to GPU) this is\na no-op.\n\nIn the Android world this corresponds to gralloc-ing with\nGRALLOC_USAGE_SW_READ | GRALLOC_USAGE_SW_WRITE.\n\nChange-Id: I52ca0ac7cc4ebfe3cb8a76aa38e437ac9a04cf02\n" }, { "commit": "5c92dea09585baf385111dad4e4e24e535a2bf0d", "tree": "f3c443e189f34a3d5ad874d2fe4c16bac1512c1e", "parents": [ "a6bdc5914a18b433d506cf1d71b5ea2804488375" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Oct 01 12:56:05 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Oct 01 12:56:05 2019 -0700" }, "message": "egl: Don\u0027t allocate texture memory for EGLImage\n\nEGLImage that has its own storage (possibly in a dmabuf) still got\nstandard texture memory allocated for it. This doubles the memory\nrequired per frame. Allocating the texture memory also isn\u0027t free,\nit can take lots of time to allocate contigous memory for a 1080p\nframe (over 25 ms), which stalls the pipeline causing frame drops.\n\nUpstream candidate.\n\nChange-Id: I78038d85992cd2742554be91f46c2010ac64595e\n" }, { "commit": "a6bdc5914a18b433d506cf1d71b5ea2804488375", "tree": "f9708f2758e982f49be90ead45855de499ce6eed", "parents": [ "c069c763e071f4908c78bb56496c1eeffeb966df" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Sep 24 10:29:57 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Oct 01 11:58:03 2019 -0700" }, "message": "Revert \"glimagesink: need add glFinish after eglswapbuffer\"\n\nTo quote the upstream maintainers:\n\"glFinish() after swapBuffers is a no-go performance wise [...]\nThe question is why do you need this in the first place ?\"\n\nhttps://gitlab.freedesktop.org/gstreamer/gst-plugins-base/issues/411\n\nglFinish stalls the entire GPU pipeline, which severely degrades\nperformance in pipelines with multiple branches, such as:\n\nglupload ! tee name\u003dt\n t. ! queue ! glfilter ! gldownload ! ...\n t. ! glimagesink\n\nglFinish at the end in glimagesink halts until any and all work\nin the pipeline context is done on the GPU including in the upper\nbranch although that isn\u0027t required. This makes both branches go\nslower. It also doesn\u0027t fix any tearing issues as that\u0027s properly\nhandled by sync fences already.\n\nThis reverts commit 397e0e785fae2da048bf9bfc3e12f572e9a26bc7.\n\nChange-Id: I6bdbe33fab36a976969e10d28e95190e47f50766\n" }, { "commit": "c069c763e071f4908c78bb56496c1eeffeb966df", "tree": "d2649d6a28819ccf73e6afd923b900965da4dc7e", "parents": [ "8caf4efce7b26173c5647dc5276c0916d5a1889b" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Sep 25 10:22:07 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Sep 27 14:40:02 2019 -0700" }, "message": "gldownload: Use EGL fences instead of glFinish to sync dmabufs\n\nNXP implemented fast texture downloads using ION backed dmabufs\nand used glFinish to ensure rendering is complete before passing\nthe CPU accessible buffer downstream. glFinish is disastrous for\nperformance though as it stalls the GPU until any and all pending\noperations are completed, as well as the single GL thread shared\nbetween al GL element in the pipeline.\n\nThis is particularly bad when there are two or more branches in\nthe pipeline (i.e. using tee). If one, say, is rendering to the\nscreen and one is doing color space conversion for inference in\nseparate paths this glFinish will force both to always wait for\ndrawing operations in the other even if they\u0027re not operating\non the same surfaces.\n\nSince the dmabuf is imported as an EGLImage we can instead use\nthe EGL_KHR_fence_sync extension to sync CPU access to the\nbuffer. Not stalling the GPU is the main benefit, but we also\ndon\u0027t have to wait for the rendering to complete in the GL\nthread so we\u0027re not blocking it either.\n\nEnd result is about 20% higher throughput in the above\ndescribed use case, as well as much smaller screen refresh rate\njitter.\n\nChange-Id: I6de5bcb2198571ac047f985d8179099595bbe8e8\n" }, { "commit": "8caf4efce7b26173c5647dc5276c0916d5a1889b", "tree": "7f0e99b780bb46f43277df6fe30c9b3e4d46e3c5", "parents": [ "3b44e50b4437050f515251cf1a9b8705ddbc4763" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Fri Oct 05 00:48:58 2018 +1000" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Tue Sep 24 10:58:24 2019 -0700" }, "message": "glimagesink: make multiview none equivalent to mono\n\nFixes the internal viewconvert to not scale buffers for output with the\nfollowing pipeline:\n\ngltestsrc ! glimagesink\n\nIt also fixes overlay composition with a resized output with an OpenGL\nupstream:\n\ngltestsrc ! timeoverlay ! glimagesink\n\nCherry picked from\nhttps://gitlab.freedesktop.org/gstreamer/gst-plugins-base/commit/6e591403a77282a7bef24d5d5c671a6d0d788c16\n\nChange-Id: Ia02f9cf89d78b96e37fe7fdddf2b33c1dcf299fc\n" }, { "commit": "3b44e50b4437050f515251cf1a9b8705ddbc4763", "tree": "a564b0f4be8817b5e35f19591105af46e0805bc8", "parents": [ "ede4f26c3530a1852fca716f386d9ba3d3b94757" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Aug 28 11:24:01 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Aug 29 10:28:58 2019 -0700" }, "message": "eglimage: Fix memory leak\n\nAlso free the GstEGLImage struct allocated by g_new0.\n\nFixes #661\n\nCherry picked from upstream master:\nhttps://gitlab.freedesktop.org/gstreamer/gst-plugins-base/commit/94984e13a9208b0b6d69dea597f42b225ef1f784\n\nChange-Id: I790e5ab89cdf3e7584a5f0a02950e8d0fd237b48\n" }, { "commit": "ede4f26c3530a1852fca716f386d9ba3d3b94757", "tree": "9ff23b60c371570e5bc30fc79cff1a6dbfd9f4d4", "parents": [ "2939bb514c72e3059ee258e9c3e99716ae0f9066" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Aug 28 09:43:42 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Wed Aug 28 10:38:39 2019 -0700" }, "message": "glupload: fix double free of EGLImage\n\nCaching EGLImage in qdata isn\u0027t thread safe nor needed. It\u0027s going away\nupstream too.\n\nhttps: //gitlab.freedesktop.org/gstreamer/gst-plugins-base/issues/375\nhttps: //gitlab.freedesktop.org/gstreamer/gst-plugins-base/issues/659\nChange-Id: I4e4a04682d53b12633ddf5dfb3bd3793b9298742\n" }, { "commit": "2939bb514c72e3059ee258e9c3e99716ae0f9066", "tree": "1863e5595f4570ed19be215468cb9b51fbac26b8", "parents": [ "3202832fee77030874b0e1408057a60649cf339a" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Aug 26 15:31:08 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Aug 26 15:31:08 2019 -0700" }, "message": "gldownload: download RGB, BGR and BGRA with dmabuf\n\nAllow these formats to be downloaded using ION backed dmabufs,\nwhich is orders of magnitude faster than PBO transfers.\n\nThis change depends on an updated imx-gpu-viv and allows for GPU\nbased conversions from YUV (camera, VPU) to inference friendly\nRGB (not RGBA) at the native source rate (30 fps).\n\nChange-Id: I04c3ca9261d9ba911dff14c0df9b996637b481fb\n" }, { "commit": "3202832fee77030874b0e1408057a60649cf339a", "tree": "46bfb0a0a9772f9107a465e40c9605cad3eacd6b", "parents": [ "eefc722e664fa19d5ba9f23cf777d858eae9370e" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Aug 26 15:19:43 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Aug 26 15:19:43 2019 -0700" }, "message": "gldownload: don\u0027t pre-allocate ion buffers\n\nNXP\u0027s gldownload fastpath requests 3 ION buffers to be created in the\npool without any upper bound. That means 3 buffers are allocated the\nmoment the pool is activated and there\u0027s no upper bound on how many\nthe pool can allocate.\n\nThis is an issue when working with larger buffers as we can easily\nrun out of CMA memory. The cost of allocating an ION buffer when\nneeded is negligible. glupload for example doesn\u0027t even use a pool.\nSince gldownload is using a pool the already negligible cost is\nhidden further.\n\nIn most common cases (i.e. downstream doesn\u0027t hold on to downloaded\nbuffers) this change results in only one buffer being allocated.\nFor a 1080p YUY -\u003e RGBA conversion that\u0027s downloaded that\u0027s a\nCMA savings of ~16 MB. For downstream elements that hold on to buffers\n(e.g. certain double buffered sinks) more buffers will be allocated as\nneeded, up to 3 to prevent OOM.\n\nChange-Id: I68172d7e9d1bf634900cb639f90843768ac525ae\n" }, { "commit": "eefc722e664fa19d5ba9f23cf777d858eae9370e", "tree": "5b1f7bbf2d084a948feabeab3df5973f1395f3cd", "parents": [ "d521bfbee8bc742707d1aa8afb2c973fde118a8c" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Aug 26 14:19:04 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Mon Aug 26 14:19:04 2019 -0700" }, "message": "Fix regression of BGRA upload color channel switched\n\nCommit 300b01b0a8b11a266f0cdd6233750fafc1e2125c ported the fix\nfrom plugins-bad commit a4b72743dd5b8ead623f4ce92a1b05ef791db67f\nbut added the call too _drm_direct_fourcc_from_info to the wrong\nfunction.\n\nChange-Id: I6051ec4f501cdde234e1503c5f0d569135f793d5\n" }, { "commit": "d521bfbee8bc742707d1aa8afb2c973fde118a8c", "tree": "870426ec628afaf5dd4b4cc7f52cff07c5b17f2f", "parents": [ "cbd278e835db0477fbf7ca815b7f3cde6b5cd0a4" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Aug 15 13:30:23 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Aug 15 16:42:47 2019 -0700" }, "message": "glupload: Fix NXP\u0027s transform_caps\n\nNXP added a fast dmabuf upload path that depends on Vivante features.\nIn this mode input is converted to RGBA backed by a single EGLImage so\nthey changed transform_caps to hard coded RGBA. They however didn\u0027t:\n\n* remove colorimetry and chroma-site from the caps which doesn\u0027t make sense\nfor a non planar format like RGBA. This leads to downstream warnings\nlike \"invalid matrix 3 for RGB format, using RGB\" to be produced for\nevery frame.\n\n* change upstream caps transform at all leading to allocation query\nmismatches. This is most obvious when a tee is present since it now\naggregates allocation queries. At least one downstream tee branch\ngets the wrong caps with subsequent all 0 pixels when frame is mapped.\n\nCherry-picked from upstream implementation of the same feature:\nhttps://github.com/GStreamer/gst-plugins-base/commit/507e31dfc00256a9cb2081aaab98a7728b0197c3\n\nChange-Id: I2e118c8559da44fd703c5071c386b7e79ff6d69a\n" }, { "commit": "cbd278e835db0477fbf7ca815b7f3cde6b5cd0a4", "tree": "8c5c945db20125beb4b6f244327d465de97e1637", "parents": [ "b916a74f1674f6226fc808a8d99e4a2f00fecce7" ], "author": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Aug 15 13:22:11 2019 -0700" }, "committer": { "name": "Jonas Larsson", "email": "ljonas@google.com", "time": "Thu Aug 15 13:22:11 2019 -0700" }, "message": "glupload: Don\u0027t prealloc any buffers in dmabuf pool\n\nThis pool logic was added by NXP. The pool gets added to any and all\nallocation queries and in certain auto plug cases the pool gets\nactivated without any buffers being dequeued before reconfig. This\nleads to memory churn and in some cases OOM. Instead allocate buffers\nonly when they\u0027re needed.\n\nChange-Id: Ic38c859db80098a2f81511dfae86fbb0330f1077\n" }, { "commit": "b916a74f1674f6226fc808a8d99e4a2f00fecce7", "tree": "4527a5878822ce92f8e8c3e11ce37872b37f886a", "parents": [ "80e1febe8ca63da591e536d2d2071bac29f4d214" ], "author": { "name": "Alex Van Damme", "email": "atv@google.com", "time": "Mon Jun 24 13:58:26 2019 -0700" }, "committer": { "name": "Alex Van Damme", "email": "atv@google.com", "time": "Mon Jun 24 13:58:26 2019 -0700" }, "message": "Demote buggy FPS prints from g_print to GST_DEBUG.\n\n- Port of 646ad50df63af358aa93109d1f5ded2edcd138d1 from\nimx-gst-plugins-bad to imx-gst-plugins-base\n\nChange-Id: I2894722961a5ec736a5cbe5618037fd5ae4bab14\n" }, { "commit": "80e1febe8ca63da591e536d2d2071bac29f4d214", "tree": "fdb463fb5f008e4b036ace0d4f9535b835a646f2", "parents": [ "300b01b0a8b11a266f0cdd6233750fafc1e2125c" ], "author": { "name": "Alex Van Damme", "email": "atv@google.com", "time": "Mon Jun 24 13:57:59 2019 -0700" }, "committer": { "name": "Alex Van Damme", "email": "atv@google.com", "time": "Mon Jun 24 13:57:59 2019 -0700" }, "message": "Use highp float precision only in YUY2 color convert shaders.\n\n- Port of e15c8166d4fcfd65c7030d53da57bd30feae7606 from\nimx-gst-plugins-bad to imx-gst-plugins-base\n\nChange-Id: I065afc9a93bbe646626566bae8f55d3191c3dfa3\n" }, { "commit": "300b01b0a8b11a266f0cdd6233750fafc1e2125c", "tree": "7203ccd8758f01a342b925e19c53fc3f458de699", "parents": [ "f62f8ab685e549eefaecdb1e700239a5a408fcad" ], "author": { "name": "Alex Van Damme", "email": "atv@google.com", "time": "Mon Jun 24 13:57:16 2019 -0700" }, "committer": { "name": "Alex Van Damme", "email": "atv@google.com", "time": "Mon Jun 24 13:57:16 2019 -0700" }, "message": "Fix NXP\u0027s custom glupload path for certain formats.\n\n- Port of a4b72743dd5b8ead623f4ce92a1b05ef791db67f from\nimx-gst-plugins-bad to imx-gst-plugins-base\n\nChange-Id: I9a21fe69cbeb1a330ac6a31b1d8e5c17cf08d7c3\n" }, { "commit": "f62f8ab685e549eefaecdb1e700239a5a408fcad", "tree": "d3a87504c9b36434d06e237231176869689d51ac", "parents": [ "cd76ab61cfeeb7b43670f2d230a9d438e06b6fde" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Wed Jan 30 15:28:08 2019 +0800" }, "committer": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Wed Jan 30 15:28:08 2019 +0800" }, "message": "MMFMWK-7783 eglimage: refine map between drm fourcc with GST format when import dmabuf\n\nwhen import dmabuf using eglimage, should take care of the map\nbetween drm fourcc with GST format. map fourcc RGBx to RGBA will\ncause GPU take x component as alpha. As a result the background is\nin black since x \u003d 0.\n\nSigned-off-by: Haihua Hu \u003cjared.hu@nxp.com\u003e\n" }, { "commit": "cd76ab61cfeeb7b43670f2d230a9d438e06b6fde", "tree": "b3a81494e458d0289dfe9634eb4575ff1c683128", "parents": [ "be30b3d6635bf7e97edb551e15e1c97ee3d6ab74" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Jan 29 17:17:54 2019 +0800" }, "committer": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Jan 29 17:17:54 2019 +0800" }, "message": "MMFMWK-8395 video-frame: correct copy width when format is NV12_10LE\n\nNV12_10LE one component occupy 10bit or 20bit which is 1.25Byte or 2.5Byte\nplane 0 10bit and plane 1 20bit (UV as one component)\ncopy width should be w \u003d (plane \u003d\u003d 0 ? 1:2) * comp_width * 5 / 4;\n\nupstream status: imx specific\nSigned-off-by: Haihua Hu \u003cjared.hu@nxp.com\u003e\n" }, { "commit": "be30b3d6635bf7e97edb551e15e1c97ee3d6ab74", "tree": "3d99c4f2612efe55ff81d37aa7aadeb825e2baaa", "parents": [ "c26e14b5ddd12f880a84d1a77ec53dbcab9e159f" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Jan 24 15:56:39 2019 +0800" }, "committer": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Jan 24 15:56:39 2019 +0800" }, "message": "MMFMWK-8402 glcolorbalance: fix pipeline crash when init\n\ncaps should be copied in gcb_transform_internal_caps,\nownership is not passed to the function\n\nupstream status: backport[1.14.4]\nhttps://gitlab.freedesktop.org/gstreamer/gst-plugins-base/issues/546\n" }, { "commit": "c26e14b5ddd12f880a84d1a77ec53dbcab9e159f", "tree": "2b41ead1c6faad645856444f0a8f356a1765b1d0", "parents": [ "a802b0b599e27454d55d49f2163a2e9251509dec" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Nov 13 21:26:01 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "MGS-4155 glupload: fix directviv internal buffer pool is not used\n\nGST_GL_HAVE_IONDMA can not indicate ion is unavailable,\nwe need try to obtain ion allocator to know whether ion\nis enabled in kernel.\n\nupstream status: imx specific\n" }, { "commit": "a802b0b599e27454d55d49f2163a2e9251509dec", "tree": "20fc299a53c6be67e0b5d402385cb896debac89c", "parents": [ "4e20437a724d9489623577011bc83cd563587dd5" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Oct 25 10:09:00 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "MMFMWK-8235 glupload: add internal physical buffer pool\n\nFor legacy platform with G2D, add internal buffer pool\nand copy non physical memory to internal buffer to enable\ndirectviv upload. Only when output format is RGBA need this,\notherwise use default glmemory to do upload.\n\nupstream status: imx specific\n" }, { "commit": "4e20437a724d9489623577011bc83cd563587dd5", "tree": "f5c4b5d84d664d6408939a92300339683309060c", "parents": [ "cf7e80c648879b86ac33566fc81b3aa7588d2774" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Mon Sep 03 10:53:34 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "enable eac3 pass-through mode\n\n- For eac3 need using 192K sample rate\n- and the sample count for alsa drvier need counted by 2byte per channel/sample\n\nSigned-off-by: Lyon Wang \u003clyon.wang@nxp.com\u003e\n" }, { "commit": "cf7e80c648879b86ac33566fc81b3aa7588d2774", "tree": "93ad824e1bb05d09086f54e53567d11753bf3800", "parents": [ "625f3164449118588d479b4f43af09b5d2d819f4" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Mar 30 13:43:05 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "MMFMWK-8218 subparse: fix critical log print out when set rate \u003c0 with external subtitle\n\nWhen playback with external subtitle, subparse will force its own start\nand stop position when seek. When rate \u003c 0, critial log will print out\nassertion \u0027stop !\u003d -1\u0027, so need avoid it when rate \u003c 0.\n\nupstream status: Pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d771648\n" }, { "commit": "625f3164449118588d479b4f43af09b5d2d819f4", "tree": "d07709177e7c5c503a5a700e8bc4acb27e6ad3ee", "parents": [ "a68ac02c06acea66ca3d50f09dd9f77da7d1327d" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Tue Sep 04 10:51:22 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "Fix build break for ion on 4.14 kernel [YOCIMX-2861]\n\nThe kernel header version check for the conditional\nion code should be testing if we have a 4.9 or 4.14 kernel.\nWhat was coded tests that the kernel version is less\nthan 4.14.34. This causes a build break for 4.14 builds\nbecause the standard libc headers are version 4.14.0.\n\nSigned-off-by: Tom Hochstein \u003ctom.hochstein@nxp.com\u003e\n" }, { "commit": "a68ac02c06acea66ca3d50f09dd9f77da7d1327d", "tree": "983576df9f8e3e6eea1b00c2c6cb9f271283a3bb", "parents": [ "46ca59b3eee471897da8b7a6cf14ef9758918782" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@freescale.com", "time": "Thu Dec 10 14:32:43 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "Disable orc optimization for lib video in plugins-base\n\n- the orc optimization for lib video in plugins base may\ncause segmentation fault\n- disalbe orc optimization for lib video and just use the c source\n\npackage: gstreamer1.0-plugins-base\nCommunity ticket: https://bugzilla.gnome.org/show_bug.cgi?id\u003d759286\nUpstream status: pending\n\nSigned-off-by: Lyon Wang \u003clyon.wang@freescale.com\u003e\n" }, { "commit": "46ca59b3eee471897da8b7a6cf14ef9758918782", "tree": "69538fe5c624aa10734558d8a98a9b7884706685", "parents": [ "8d08d4be1882bdc8cf5f0ad41242d953de6148c7" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Aug 21 13:20:29 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "MMFMWK-8190 glupload: refine transform caps of directviv uploader\n\nif downstream only accept YUV format (eg, add capsfilter), then\nwe should passthrough color format and don\u0027t extend to directviv\nsupport format\n\nupstream status: Pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d797002\n" }, { "commit": "8d08d4be1882bdc8cf5f0ad41242d953de6148c7", "tree": "2c1edd73f3d2ad0cfe93fea01952500513f5a712", "parents": [ "c652ead787c076be9734dd29e21001818997df70" ], "author": { "name": "Song Bing", "email": "bing.song@nxp.com", "time": "Wed Aug 15 14:51:09 2018 -0700" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "gstimxcommon: Enable VPU feature for i.MX8QM B0\n\nupstream status: imx specific\n\nSigned-off-by: Song Bing \u003cbing.song@nxp.com\u003e\n" }, { "commit": "c652ead787c076be9734dd29e21001818997df70", "tree": "7172838262978a17de5cc80d8f14f26f4bef3370", "parents": [ "b28731f3140934e797cc4936472957bc20f3836b" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Aug 02 16:14:03 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "gstimxcommon: add some platform marcos\n\nupstream status: imx specific\n\nSigned-off-by: Haihua Hu \u003cjared.hu@nxp.com\u003e\n" }, { "commit": "b28731f3140934e797cc4936472957bc20f3836b", "tree": "91c7df8a843c9d3f1ad64786e40848c1ae7078dc", "parents": [ "bb3530c1da59ee7c300f069b34691a647c2a917c" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Jul 24 14:05:17 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "MMFMWK-8077 glupload: respect to downstream format priority\n\nMake upstream caps format respect to downstream\u0027s priority\nto try hard to choose RGBA format if possible\n\nupstream status: imx specific\n" }, { "commit": "bb3530c1da59ee7c300f069b34691a647c2a917c", "tree": "b2137b7380366778566954a875b45efcb7a33d04", "parents": [ "4b3e37f9b7d7b1c011f4878a5f274d2b2aaff899" ], "author": { "name": "Song Bing", "email": "bing.song@nxp.com", "time": "Tue Jul 10 17:50:29 2018 -0700" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "MMFMWK-8113 [ion] Enable ion allocator based on Kernel 4.14\n\nion driver changed. Enable ion allocator based on Kernel 4.14.\n\nSigned-off-by: Song Bing \u003cbing.song@nxp.com\u003e\n" }, { "commit": "4b3e37f9b7d7b1c011f4878a5f274d2b2aaff899", "tree": "e05f972afefdcb0408a29af19c9f0c7f8d7b7d45", "parents": [ "0fe8c176dd264075958d2ffea428c241d291821d" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Fri Jul 06 13:18:54 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "alsasink: Add passthrough property\n\n- Add property pass-through to control\n if enable pass-through mode\n\nSigned-off-by: Lyon Wang \u003clyon.wang@nxp.com\u003e\n" }, { "commit": "0fe8c176dd264075958d2ffea428c241d291821d", "tree": "68d6a73734796a23a8729d3bef58c67e01185011", "parents": [ "cf1b1856ce87f64b4e4b8c79a0d3b78671991ff4" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Thu Jun 21 15:32:04 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "Open iec958 audio device\n\n- Our audio driver not support open device\nwith AES parameter, so just use hw: device name\n- i.mx specific for audio passthrough\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d757258\n\nSigned-off-by: Lyon Wang \u003clyon.wang@nxp.com\u003e\n" }, { "commit": "cf1b1856ce87f64b4e4b8c79a0d3b78671991ff4", "tree": "0e9de27441577428426c894b69966134863e22f5", "parents": [ "2b29b1b7817b7ea197119b8085dec9ed15c46d7f" ], "author": { "name": "Arnaud Pouliquen", "email": "arnaud.pouliquen@st.com", "time": "Thu Nov 05 18:47:19 2015 +0100" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:33 2018 +0800" }, "message": "iec61937: set iec958 boolean before use it\n\nDuring prepare alsa-\u003eiec958 is tested before set.\nEffect is that alsa device is not opened in iec958 mode,\nso AES value are not transmitted to alsa driver.\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d757258\n\nSigned-off-by: Arnaud Pouliquen \u003carnaud.pouliquen@st.com\u003e\n" }, { "commit": "2b29b1b7817b7ea197119b8085dec9ed15c46d7f", "tree": "ff82905dea6d8578d300a6e98ae8d281c74e7437", "parents": [ "90a66834d06ab3b382da764320694ba19d82e6c0" ], "author": { "name": "Arnaud Pouliquen", "email": "arnaud.pouliquen@st.com", "time": "Fri Nov 06 10:01:50 2015 +0100" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "iec61937: force hw_param channels to stereo\n\nIn case of pass-through, iec frames are sent over a stereo PCM\nstream. So alsa device should be configured with channels \u003d 2.\nThis is needed to be compatible with SPDIF electrical/optical output\nthat supports only stereo in PCM mode and up to 8 channels in iec61937 mode.\n\nSigned-off-by: Arnaud Pouliquen \u003carnaud.pouliquen@st.com\u003e\n" }, { "commit": "90a66834d06ab3b382da764320694ba19d82e6c0", "tree": "e2d6aecf67283a514943ec255147cd947fadf3a3", "parents": [ "a34468be251df3ae6f2b2fbf17b4200d9941744a" ], "author": { "name": "Arnaud Pouliquen", "email": "arnaud.pouliquen@st.com", "time": "Tue Nov 03 14:13:10 2015 +0100" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "alsasink: fix iec958 format detection\n\nAdd specific function that first close the current pcm device\nbefore trying to re-open it with AES parameters.\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d757258\n\nSigned-off-by: Arnaud Pouliquen \u003carnaud.pouliquen@st.com\u003e\n" }, { "commit": "a34468be251df3ae6f2b2fbf17b4200d9941744a", "tree": "8f8ac753d505b33c78e54e5508989f2a5bf51af1", "parents": [ "3f201101290ede2c02f5a7e2718d4d207d6f6f3d" ], "author": { "name": "Arnaud Pouliquen", "email": "arnaud.pouliquen@st.com", "time": "Fri Oct 30 09:45:46 2015 +0100" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "iec61937: incoherent endianness in payload\n\nOrder header based on specified endianness and not system endianness\nFix following bugzilla\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d678021\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d757258\n\nSigned-off-by: Arnaud Pouliquen \u003carnaud.pouliquen@st.com\u003e\n" }, { "commit": "3f201101290ede2c02f5a7e2718d4d207d6f6f3d", "tree": "f96c9f36884bf467e489ee45e47ac2de0d21fd51", "parents": [ "3df2e0664ce20bd854cafdfc5b87a9449f1d23de" ], "author": { "name": "Arnaud Pouliquen", "email": "arnaud.pouliquen@st.com", "time": "Fri Oct 30 09:45:27 2015 +0100" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "alsasink: endianess fix for iec61937\n\nif format supported by driver is S16_LE, data should be swapped only for big endian systems\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d757258\n\nSigned-off-by: Arnaud Pouliquen \u003carnaud.pouliquen@st.com\u003e\n" }, { "commit": "3df2e0664ce20bd854cafdfc5b87a9449f1d23de", "tree": "fc2db86e69b30a05a2070991e78d8d58ac7376b1", "parents": [ "23195eaad7a5fba5ad4d43839b0d3409671b4190" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu May 10 10:58:12 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "MMFMWK-7996 gstimxcommon: add marcos for mscale 845s chip code\n\nupstream status: imx specific\n" }, { "commit": "23195eaad7a5fba5ad4d43839b0d3409671b4190", "tree": "9ecd6d69c11c706b271ff026ed54509c66f54959", "parents": [ "c12ede3183e8f29bb6fe2f1f1d6ee8618aa9f753" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Thu May 03 17:28:02 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "[MMFMWK-7988] Add i.MX8MM featuer list\n\n- Add i.MX8MM feature list accoridng to soc_id\n\nSigned-off-by: Lyon Wang \u003clyon.wang@nxp.com\u003e\n" }, { "commit": "c12ede3183e8f29bb6fe2f1f1d6ee8618aa9f753", "tree": "b642033f3b2f913e8c519228f4454460d8703a06", "parents": [ "7729f55ba70bbd5d86365a1ac882c7c9ebcec781" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Apr 26 13:01:05 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "MMFMWK-7960 viv-fb: export viv-fb display api\n\nqmlgl plugin will use this api\n\nupstream status: backport [1.15.1]\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d795562\n" }, { "commit": "7729f55ba70bbd5d86365a1ac882c7c9ebcec781", "tree": "9b8b450463d43c7e25c7f535e438b481b43c6871", "parents": [ "397e0e785fae2da048bf9bfc3e12f572e9a26bc7" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Apr 24 13:48:53 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "MMFMWK-7960 viv-fb: install gstgldisplay_viv_fb.h into build result\n\ngstgldisplay_viv_fb.h is need by qmlgl plugin in gst-good package\n\nupstream status: pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d795499\n" }, { "commit": "397e0e785fae2da048bf9bfc3e12f572e9a26bc7", "tree": "cadac359004a5e25902059204c04c4238d3b58f6", "parents": [ "6a5f96dda4e8c52c9b55146191fbdc2d75cf90c0" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Dec 28 16:37:20 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "glimagesink: need add glFinish after eglswapbuffer\n\nwhen video playback using glimagesink, video tearing occurs. eglswapbuffer\nis an async call that is to say the buffer maybe still used by GPU\nbut we are already prepare next buffer.\n\nupstream status: pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d791937\n" }, { "commit": "6a5f96dda4e8c52c9b55146191fbdc2d75cf90c0", "tree": "d453a238cb251a2b39f703ebca5bb235cbced3fb", "parents": [ "b57538230d2c20a9c5f1a0d6dea55a0cd7c2cbd4" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Apr 12 16:22:25 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "gloverlaycompositor: add subtitle rotate support follow video\n\nupstream status: Pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d790470\n" }, { "commit": "b57538230d2c20a9c5f1a0d6dea55a0cd7c2cbd4", "tree": "0c419c944ae5a692395c119ba35261713ef746eb", "parents": [ "e2484b4ffce620a9a574ae637b29f97c6eeeaaa6" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Fri Nov 25 14:48:44 2016 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "Add ion dmabuf support in gldownload\n\nSupport copy into dma-fb buffer if support the buffer format\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Haihua Hu \u003cjared.hu@nxp.com\u003e\n" }, { "commit": "e2484b4ffce620a9a574ae637b29f97c6eeeaaa6", "tree": "6392b60d5453dd972da9c380adebf97865fe4d61", "parents": [ "aafbe5b5f936b56ef72e784331315e0d899efa43" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Wed Apr 11 18:02:04 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "glupload: add ion dmabuf support in glupload\n\n1. Support one texture for YUV format in dmabuf uploader\n2. Propose ion dma-fd buffer pool to upstream to avoid memory copy\n3. If upstream don\u0027t chose the proposed buffer pool, then create\n our own and do copy to avoid memory copy from CPU to GPU side\n4. Add buffer alignmentw\n\nUpstream Status: Inappropriate [i.MX specific]\n" }, { "commit": "aafbe5b5f936b56ef72e784331315e0d899efa43", "tree": "2a80273249c8d37ad3d53cb869c43a3747ede24d", "parents": [ "825b0921e22c766214af71214b91f7b58523d4e5" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Wed Apr 11 14:54:52 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "Accelerate gldownload with directviv\n\n1.Propose a physical buffer pool to upstream in gldownload\n2.Bind the physical buffer with texture via dirctviv\n3.In gldownload, wrap the physical buffer to gstbuffer, pass to\n downstream plugins.\n4.Add some configure check for g2d and phymem\n\nFrom GST-1.14, add GST_GL_API ahead api to export symbol\nadd include \"gstglfuncs.h\" when need call gl api\n\nUpstream-Status: Inappropriate [i.MX specific]\n" }, { "commit": "825b0921e22c766214af71214b91f7b58523d4e5", "tree": "6bdc7ed2f3935c0c8b4d62878503899ff4a6222a", "parents": [ "3e1d5364d266e5084292c3491c7642f654cc6441" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Apr 11 16:55:11 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "Update IMX_GST_PLUGIN_DEFINE define for GST1.14\n\n- As macro GST_PLUGIN_DEFINE is upgraded in GST1.14\n(change C structure into function\ncommit e7ede5a487, https://bugzilla.gnome.org/show_bug.cgi?id\u003d779344)\n\nSo our macor cannot use \".\" in name field\n\n-- i.MX specific\n\nSigned-off-by: Lyon Wang \u003clyon.wang@nxp.com\u003e\n" }, { "commit": "3e1d5364d266e5084292c3491c7642f654cc6441", "tree": "cd7afc97b6ee7004c2fe27c56077035c762a9623", "parents": [ "641863d769f7ab1878dde778935d7985fa7e42d7" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Mon Apr 09 15:58:52 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "fix glplugin not build in wayland backend\n\nupstream status: imx specific\n" }, { "commit": "641863d769f7ab1878dde778935d7985fa7e42d7", "tree": "3e6b17633f79f6cc7edf0ca8d3f8d2121d9ceb94", "parents": [ "1cdd2b9a3ebdc4db87d676fdb6f63f97674236e3" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Tue Apr 10 16:06:14 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "Add imx physical memory allocator\n\nUpstream status: [i.MX specific]\n\nSigned-off-by: Haihua Hu \u003cjared.hu@nxp.com\u003e\n" }, { "commit": "1cdd2b9a3ebdc4db87d676fdb6f63f97674236e3", "tree": "eb76f16e37705a27c08a31c61894e862c1848805", "parents": [ "e271f127ba0b79dd45d82c76078e7d0afe730785" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Mon Apr 02 16:20:01 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "MMFMWK-7895 [MX8MQ] Enable Hantro video tiled and\n\n Move physical meta to plugin bad.\n move to base since 1.14\n\n Signed-off-by: Song Bing \u003cbing.song@nxp.com\u003e\n" }, { "commit": "e271f127ba0b79dd45d82c76078e7d0afe730785", "tree": "77d3694b38b15aeb8950ecf047bc4af8c70a2ce9", "parents": [ "4d67853b574f12bce7d9b0e02f36f8cce1166416" ], "author": { "name": "Song Bing", "email": "bing.song@nxp.com", "time": "Wed Sep 13 13:39:53 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "ionmemory: support get phys memory\n\nUpstream Status: Pending\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d768794\n" }, { "commit": "4d67853b574f12bce7d9b0e02f36f8cce1166416", "tree": "3150428495c0792b4f968ea9d0fe43bbdb85d790", "parents": [ "de9401f07a12be7cb7f20a78c636953d731065d8" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Mon Apr 02 15:51:48 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "ionmemory: dmabuf memory allocator based on ion driver.\n\nUpstream Status: Pending\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d768794\n" }, { "commit": "de9401f07a12be7cb7f20a78c636953d731065d8", "tree": "d7d6d16ea02bf519c231f27ba744abfbf8ad0e5a", "parents": [ "87c28842a8bce3e957c0321fb687d85e3a84acb7" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Wed Sep 13 16:42:21 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "glupload: add crop meta support in dmafd uploader\n\nget video crop meta from input buffer and update video info\n\nupstream status: pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d787616\n" }, { "commit": "87c28842a8bce3e957c0321fb687d85e3a84acb7", "tree": "1c23f699c4541833856401d020a4c577235a3035", "parents": [ "73536bbb081d3ae08f46a4e51b3281cc51d46dac" ], "author": { "name": "Song Bing", "email": "bing.song@nxp.com", "time": "Fri Mar 23 15:13:55 2018 -0700" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "allocators: Add GstDmabufMeta\n\nThis meta can be added to the allocation query to indicate support\nfor dmabuf memory. Optionally the drm_modifier can also be set in\nthe allocation query to indicate that the element supports some\ncertain set of drm modifiers (swizzling/tiling/compression/etc).\nWhen the meta is attached to a buffer, the drm_modifier should\nonly be one which the src and sink both indicated support for in\nthe allocation query.\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d779146\n" }, { "commit": "73536bbb081d3ae08f46a4e51b3281cc51d46dac", "tree": "2b31bd5ec2d490cb417253dc6142f1d7b9be4361", "parents": [ "abff9ca882a3de53341283da704a74258cc052b8" ], "author": { "name": "Haihua Hu", "email": "b55597@freescale.com", "time": "Tue Dec 08 16:06:34 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:32 2018 +0800" }, "message": "Fix glimagesink wayland resize showed blurred screen\n\nFor imx, wl_egl_window type is not a wl_proxy object. Can not set\nqueue to wl_egl_window object.\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Haihua Hu \u003cb55597@freescale.com\u003e\n" }, { "commit": "abff9ca882a3de53341283da704a74258cc052b8", "tree": "563a958df2d7084e760a99bb8da23f62f36932d8", "parents": [ "0ab66a830add68f1ad0fb8173c1859e41ca405ba" ], "author": { "name": "Haihua Hu", "email": "b55597@freescale.com", "time": "Thu Feb 25 13:53:20 2016 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "glcolorconvert: convert YUV to RGB use directviv\n\nAdd a property \"disable_passthrough\" in glcolorconvert for enable/disable passthrough.\nWhen need convert YUV to RGB with directviv, set it to be TRUE.\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Haihua Hu \u003cb55597@freescale.com\u003e\n\nConflicts:\n\tgst-libs/gst/gl/gstglcolorconvert.c\n" }, { "commit": "0ab66a830add68f1ad0fb8173c1859e41ca405ba", "tree": "720157e1b4cb506def77cdaf8b46f89093297e72", "parents": [ "4719288e1848656f23d70da7a1d09a3171e1340a" ], "author": { "name": "Haihua Hu", "email": "b55597@freescale.com", "time": "Wed Nov 18 15:10:22 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "Add fps print in glimagesink\n\nIn GST-1.6, Pipeline will set start time to 0 when state\nchange form PAUSE to READY, so get start time in state change\nPLAYING_PAUSE.\n\nUpstream-Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Haihua Hu \u003cb55597@freescale.com\u003e\n\nConflicts:\n\text/gl/gstglimagesink.h\n" }, { "commit": "4719288e1848656f23d70da7a1d09a3171e1340a", "tree": "136df30d02ede9a7630e3e31b8cbb523a147c146", "parents": [ "f27f2a4cd443449291901aa045d4746a8eaad946" ], "author": { "name": "Haihua Hu", "email": "b55597@freescale.com", "time": "Fri Nov 13 10:51:25 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "support video crop for glimagesink\n\n1.Add video crop meta copy in glupload\n2.Calculate the new texture coordinate in vertices array and bind to buffer object\n3.Make glimagesink only updating vertices array when video crop meta changed\n\nUpstream-Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Haihua Hu \u003cb55597@freescale.com\u003e\n" }, { "commit": "f27f2a4cd443449291901aa045d4746a8eaad946", "tree": "8e814e9830509e039d9b9d44e36b0b066c5d8c5c", "parents": [ "0a10c8e465b844571c14d8fb776fc2d5aa214863" ], "author": { "name": "Jian", "email": "Jian.Li@freescale.com", "time": "Tue Nov 17 14:39:07 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "Change wayland default res to 1024x768\n\nChange wayland backend default window resolution\nfrom 320x240 to 1024x768\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Jian \u003cJian.Li@freescale.com\u003e\n" }, { "commit": "0a10c8e465b844571c14d8fb776fc2d5aa214863", "tree": "9113bccaab9d498d15a9b5a39b17bb4c906a45c0", "parents": [ "60d37a2ee2a3bce08ac422bdb50063b3aaab15e0" ], "author": { "name": "Song Bing", "email": "b06498@freescale.com", "time": "Mon Jun 08 17:06:22 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "glfilter: Lost frame rate info when fixate caps\n\nLost frame rate info when fixate caps. It will cause\ndown stream element fail, such avimux.\n\nUpstream Status: Waiting for review.\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d750545\n" }, { "commit": "60d37a2ee2a3bce08ac422bdb50063b3aaab15e0", "tree": "574cb8ce0ab9e5eaedce89e0e23ebe7394cb9113", "parents": [ "b157e179a3d9d11f289835382b2ad499f66117f1" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Thu Feb 08 15:26:06 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "Move gstimxcommon.h into gst-plugins-base\n\nMMFMWK-7883\n\nMove gstimxcommon.h into base for future imx specified dev\n\nSigned-off-by: Lyon Wang \u003clyon.wang@nxp.com\u003e\n" }, { "commit": "b157e179a3d9d11f289835382b2ad499f66117f1", "tree": "777f586083146f4956ff34d91d742d8086315f2b", "parents": [ "46456ace160dab6e4e6de5cfe89fdb7546f2feb3" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Tue Jan 30 11:04:31 2018 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "add fullRange to hdr10 metadata\n\nupstream status: i.Mx specific\n" }, { "commit": "46456ace160dab6e4e6de5cfe89fdb7546f2feb3", "tree": "b050b7099d8809acd047a5d73e8eb157a5e122d9", "parents": [ "10eb88d00ee82269a3748716d79be54b0c9418a3" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Nov 30 17:43:20 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "basetextoverlay: need avoid idx exceed memory block number\n\nwhen check whether video buffer is read only, the gst_buffer_get_memory call\nshould make sure idx don\u0027t exceed the total memory block number\n\nupstream status: i.mx specific\n\nSigned-off-by: Haihua Hu \u003cjared.hu@nxp.com\u003e\n" }, { "commit": "10eb88d00ee82269a3748716d79be54b0c9418a3", "tree": "79fe4c36e3f192b9cedef8a346d5b35c457c7545", "parents": [ "8124149a5af8a3940a6affff20b67984c9b50bdb" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Mon Nov 06 15:05:47 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "fbmemory: need unmap if mapping flags are not subset of previous\n\nupstream status: Pending\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d789952\n" }, { "commit": "8124149a5af8a3940a6affff20b67984c9b50bdb", "tree": "5a6e0f1160bc1b11c9cd985f5128fd55b935b1c5", "parents": [ "7ad5590b6eed6bbec957a15b7b67d5ec177a10f7" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Thu Nov 02 10:43:49 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "Add 10bit packed NV12 format\n" }, { "commit": "7ad5590b6eed6bbec957a15b7b67d5ec177a10f7", "tree": "ffd44b70998b5fa39516cf109129b6d1e25e9221", "parents": [ "cbf8d048f0c0b22ed63b7f0628e4fb4d3504d692" ], "author": { "name": "Haihua Hu", "email": "jared.hu@nxp.com", "time": "Wed Nov 01 14:22:59 2017 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "add HDR10 metadata type\n" }, { "commit": "cbf8d048f0c0b22ed63b7f0628e4fb4d3504d692", "tree": "0d7d3adbe4c8b2d804a30c6b6b665511f42bf2cd", "parents": [ "a9d4e7a3a18c060355dc8817bbb0584a4f687c7f" ], "author": { "name": "Song Bing", "email": "bing.song@nxp.com", "time": "Wed Sep 13 13:37:17 2017 -0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "dmabuf: set fd memory to keep mapped\n\nset fd memory to keep mapped.\n\nUpstream Status: Pending\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d768794\n" }, { "commit": "a9d4e7a3a18c060355dc8817bbb0584a4f687c7f", "tree": "c59d9156d2457d75e8e8462e05cc547c185fa55c", "parents": [ "87a1fa1d082714be7dcb549d6fd7cda51b53b925" ], "author": { "name": "Song Bing", "email": "b06498@freescale.com", "time": "Mon Jan 11 14:51:17 2016 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "MMFMWK-7030 [Linux_MX6QP_ARD]IMXCameraApp:When Enabled \"save time to image\" item, preview, find the time can not display completely. 100%\n\nAs IPU need 8 pixels alignment, add one workaround in base text overlay\nto generate 8 pixels alignment text video buffer. The side effect should\ncause all text a little smaller.\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Song Bing b06498@freescale.com\n" }, { "commit": "87a1fa1d082714be7dcb549d6fd7cda51b53b925", "tree": "142d432858c40ae8b3d117d38945d2cb88b78c1b", "parents": [ "1e96e05449cde418f0368fef92cd23b9db4ab59c" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@freescale.com", "time": "Tue Nov 17 14:56:47 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "gstaudiobasesink print warning istead of return ERROR.\n\nFor those clips with corrupt audio track,\nthere might be no output from audio decoder\nand thus the audio track have no chance to negotiate.\nWe can just print error warning instead of return ERROR,\nso that other track can be played normally\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d758215\n\nUpstream Status: Pending\n\nSigned-off-by: Lyon Wang \u003clyon.wang@freescale.com\u003e\n" }, { "commit": "1e96e05449cde418f0368fef92cd23b9db4ab59c", "tree": "fca94d06a50cf1981460e7f0dba6cbad670556ce", "parents": [ "f4205f77ce37d7ea6ba2a9f404678eabd22e246a" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@freescale.com", "time": "Mon Dec 15 16:52:07 2014 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "handle audio/video decoder error\n\nWhen there is input data and no output data to the end of the stream, it will\nsend GST_ELEMENT_ERROR, So the clips playing will quit.\nHowever, if only one of the tracks is corrupt, there is no need to quit other\ntracks playing.\n\nThe patch comments the GST_ELEMENT_ERROR() and just add GST_ERROR_OBJECT()\ninformation instead.\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d741542\n\nUpstream Status: Pending\n\nSigned-off-by: Lyon Wang \u003clyon.wang@freescale.com\u003e\n" }, { "commit": "f4205f77ce37d7ea6ba2a9f404678eabd22e246a", "tree": "b52a637381c3385557c9ef87c798c4121cd95add", "parents": [ "6da48a05eae8b0291a8c455a1bff0ab0b5d87c55" ], "author": { "name": "Lyon Wang", "email": "lyon.wang@freescale.com", "time": "Wed Oct 21 16:35:43 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "taglist not send to down stream if all the frame corrupted\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d737246\n\nUpstream status: Pending\n\nSigned-off-by: Jian Li \u003clj.qfy.sh@gmail.com\u003e\n" }, { "commit": "6da48a05eae8b0291a8c455a1bff0ab0b5d87c55", "tree": "7b898b398f0a4d5c89305a477ff692af2b2ed625", "parents": [ "fee098b9c0c8311eedae95e4024ce31f690b749e" ], "author": { "name": "Jian Li", "email": "jian.li@freescale.com", "time": "Mon Jun 23 14:14:07 2014 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:31 2018 +0800" }, "message": "gstplaybin remove default deinterlace flag\n\n- remove default deinterlace flag in playbin for i.MX SoCs\n\nUpstream Status: Inappropriate [platform specific]\n\nSigned-off-by: Jian Li \u003cjian.li@freescale.com\u003e\n" }, { "commit": "fee098b9c0c8311eedae95e4024ce31f690b749e", "tree": "e05a16bd54ff8a8e4494d314ce38adf0365e7a49", "parents": [ "c1158c222334ede2819f71f52de7c59786895f3f" ], "author": { "name": "zhouming", "email": "b42586@freescale.com", "time": "Wed May 14 10:16:20 2014 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:30 2018 +0800" }, "message": "ENGR00312515: get caps from src pad when query caps\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d728312\n\nUpstream Status: Pending\n\nSigned-off-by: zhouming \u003cb42586@freescale.com\u003e\n" }, { "commit": "c1158c222334ede2819f71f52de7c59786895f3f", "tree": "bf813ea8a3974860ecd62ebb217d9fa2a43be9ba", "parents": [ "85e5ff562c6024c548478bdc5be8efd428bf2c42" ], "author": { "name": "Mingke Wang", "email": "mingke.wang@freescale.com", "time": "Thu Mar 19 14:20:26 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:30 2018 +0800" }, "message": "subparse: set need_segment after sink pad received GST_EVENT_SEGMENT\n\nsubparse works in push mode, chain funciton will be called once\nup stream element finished the seeking and flushing.\nif set need_segment flag in src pad event handler, the segment\nevent will be pushed earlier, result in the subtitle text will\nbe send out to down stream from the beginning.\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Mingke Wang \u003cmingke.wang@freescale.com\u003e\n" }, { "commit": "85e5ff562c6024c548478bdc5be8efd428bf2c42", "tree": "cafc5067fd158cb2ff4bb6a27871f06b8718df4c", "parents": [ "cf08a3efaa6c3512924ade924ce89150cde9270a" ], "author": { "name": "Mingke Wang", "email": "mingke.wang@freescale.com", "time": "Thu Mar 19 14:17:10 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:30 2018 +0800" }, "message": "ssaparse: enhance SSA text lines parsing.\n\nsome parser will pass in the original ssa text line which starts with \"Dialog:\"\nand there\u0027s are maybe multiple Dialog lines in one input buffer.\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Mingke Wang \u003cmingke.wang@freescale.com\u003e\n" }, { "commit": "cf08a3efaa6c3512924ade924ce89150cde9270a", "tree": "c0fabe4122412bf632db1b4e1881425871948d07", "parents": [ "50cc8b8423dd3d20b69e82b912fc28f8fc872942" ], "author": { "name": "Mingke Wang", "email": "mingke.wang@freescale.com", "time": "Thu Mar 19 14:15:25 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:30 2018 +0800" }, "message": "gstplaysink: don\u0027t set async of custom text-sink to false\n\nset async to false lead to A/V sync problem when seeking.\nthe preroll need use GAP event instead of set async to false.\n\nUpstream Status: Inappropriate [i.MX specific]\n\nSigned-off-by: Mingke Wang \u003cmingke.wang@freescale.com\u003e\n" }, { "commit": "50cc8b8423dd3d20b69e82b912fc28f8fc872942", "tree": "a88fdcb54b47aaaa38ce5714f4e4a1c7f4d548f9", "parents": [ "384ff7d0268c71e76f7328a9ca6dfe96e4a3ab23" ], "author": { "name": "Mingke Wang", "email": "mingke.wang@freescale.com", "time": "Fri Oct 16 19:31:32 2015 +0800" }, "committer": { "name": "Lyon Wang", "email": "lyon.wang@nxp.com", "time": "Wed Dec 19 10:50:30 2018 +0800" }, "message": "basetextoverlay: make memory copy when video buffer\u0027s memory is ready only\n\n1. since gst_buffer_make_writable just lookup the refcount to determine if\n a buffer is writable, and it will use _gst_buffer_copy() which don\u0027t\n perform a deep memory copy even if the flag of a memory is set to\n GST_MEMORY_FLAG_READONLY. So, we detect the memory flag and use\n gst_buffer_copy_region with GST_BUFFER_COPY_DEEP parameter to perform\n deep memory copy. if the allocator of a memory don\u0027t support mem_copy\n interface, the it will return NULL, if this case, we can use\n gst_buffer_make_writable() to get a shared memory buffer or the orignal\n buffer if the buffer\u0027s refcount is 1.\n\nSigned-off-by: Mingke Wang \u003cmingke.wang@freescale.com\u003e\n" }, { "commit": "384ff7d0268c71e76f7328a9ca6dfe96e4a3ab23", "tree": "1c86e4d9f69b4998845e8c39163c1ebb3c992462", "parents": [ "4d89b594c332e412d56ce322d67bc9eb052c054c" ], "author": { "name": "Tim-Philipp Müller", "email": "tim@centricular.com", "time": "Tue Oct 02 22:58:06 2018 +0100" }, "committer": { "name": "Tim-Philipp Müller", "email": "tim@centricular.com", "time": "Tue Oct 02 22:58:08 2018 +0100" }, "message": "Release 1.14.4\n" }, { "commit": "4d89b594c332e412d56ce322d67bc9eb052c054c", "tree": "39611300c9ce3647191c4825b0d20feef9a41308", "parents": [ "6e30040809c9ebc47f42a8c578ce0a3ed7b84855" ], "author": { "name": "Tim-Philipp Müller", "email": "tim@centricular.com", "time": "Tue Oct 02 22:58:06 2018 +0100" }, "committer": { "name": "Tim-Philipp Müller", "email": "tim@centricular.com", "time": "Tue Oct 02 22:58:06 2018 +0100" }, "message": "Update docs\n" }, { "commit": "6e30040809c9ebc47f42a8c578ce0a3ed7b84855", "tree": "2e4a644ea1bc683bc3b410029d44c831fcb7e675", "parents": [ "54e21b7bc279d234c0cb700f7d1f3baefb3ccadb" ], "author": { "name": "Matthew Waters", "email": "matthew@centricular.com", "time": "Wed Sep 26 13:22:14 2018 +1000" }, "committer": { "name": "Sebastian Dröge", "email": "sebastian@centricular.com", "time": "Thu Sep 27 22:33:49 2018 +0300" }, "message": "glviewconvert: wait and set the gl sync meta on buffers\n\nThis may be a cause of out-of-place frames when transforming multiview\nbuffers.\n" }, { "commit": "54e21b7bc279d234c0cb700f7d1f3baefb3ccadb", "tree": "f71ef390d515885eb02e6331b197ae494f5b7080", "parents": [ "83e1bee3f10c351076e3534ee9bb084519ebb883" ], "author": { "name": "Edward Hervey", "email": "edward@centricular.com", "time": "Thu Sep 27 15:03:10 2018 +0200" }, "committer": { "name": "Sebastian Dröge", "email": "sebastian@centricular.com", "time": "Thu Sep 27 22:33:38 2018 +0300" }, "message": "tagdemux: Use upstream GST_EVENT_STREAM_START if present\n\nIf we have an upstream GST_EVENT_STREAM_START, use that one instead\nof creating a new one which could be completely different from the\nupstream one and drop information (like the stream flags and stream\nobject).\n\nOnly create a new event if we don\u0027t already have one from upstream\n\nhttps://bugzilla.gnome.org/show_bug.cgi?id\u003d797215\n" }, { "commit": "83e1bee3f10c351076e3534ee9bb084519ebb883", "tree": "ab7196c51157af94fb6f9c3c908c0ec2293f455f", "parents": [ "24c20d4ced940c64f27a62bc109e7b7360bc6b05" ], "author": { "name": "Sebastian Dröge", "email": "sebastian@centricular.com", "time": "Tue Sep 25 16:56:07 2018 +0300" }, "committer": { "name": "Sebastian Dröge", "email": "sebastian@centricular.com", "time": "Thu Sep 27 22:33:31 2018 +0300" }, "message": "glcolorconvert: Don\u0027t copy overlay composition meta over to NULL outbufs\n" } ], "next": "24c20d4ced940c64f27a62bc109e7b7360bc6b05" }