{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":629462509,"defaultBranch":"lineage-20","name":"android_hardware_samsung_slsi-linaro_graphics","ownerLogin":"K9100ii","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2023-04-18T11:13:35.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/55348214?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1705449873.0","currentOid":""},"activityList":{"items":[{"before":"f2d2cf5b414a77b58a1f82058d2c8d0f32eeb751","after":"656fa32f5ab612a1646c2f16fd094407360b6610","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-06-25T08:34:38.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"f2d2cf5b414a77b58a1f82058d2c8d0f32eeb751","after":"656fa32f5ab612a1646c2f16fd094407360b6610","ref":"refs/heads/lineage-21","pushedAt":"2024-06-25T08:34:30.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","after":"f2d2cf5b414a77b58a1f82058d2c8d0f32eeb751","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-05-22T14:31:00.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","after":"f2d2cf5b414a77b58a1f82058d2c8d0f32eeb751","ref":"refs/heads/lineage-20","pushedAt":"2024-05-22T14:30:51.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","after":"f2d2cf5b414a77b58a1f82058d2c8d0f32eeb751","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-05-22T14:30:10.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","after":"f2d2cf5b414a77b58a1f82058d2c8d0f32eeb751","ref":"refs/heads/lineage-21","pushedAt":"2024-05-22T14:29:53.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"8e8a2e09f9463bb0498c8c5ac756626f449e4fd2","after":"898bb2dc9527049554116b7bb29fa29c48f2ac93","ref":"refs/heads/lineage-20_10-e850-96","pushedAt":"2024-03-30T21:19:54.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: Fix large mmap() and munmap() length mismatches\n\nThe size of struct ExynosVideoMeta, used to determine lengths of\nmappings in mmap() calls, is very large compared to the length used in\ncorresponding munmap() calls - 64 bytes. This causes quite some memory\nleakage.\n\nUnder these BSP sources, ExynosVideoMeta has a size of 400 bytes. This\nmeans that, on every time mmap() and munmap() are called with their\nmismatched sizes in libhwc1, 400-64=336 bytes are undesirably leaked.\nWhile being very problematic on its own, in theory for real usage, that\ncould eventually cause problems particularly after many days of usage\nwith very heavy video playback (possibly even encountered by someone out\nthere, but unreported or ignored as a 'one-off thing'; no problems were\nknown before with these BSP sources).\n\nFix this problem by matching lengths used for the mmap() and munmap()\ncalls.\n\nAlso, don't map the entire size of the ExynosVideoMeta struct - Since\nonly the 'eType' member in it is checked in between the mmap() and\nmunmap() calls, and it's right at the start of the struct, the size of\nonly the 'eType' member can be mapped (and unmapped), which is only a\ntiny 4 bytes and seems to work fine.\n\nFor some trivia, this all looks like an oversight from someone at\nSamsung. In some really old BSP sources [1], mmap(), as well as\nmunmap(), is called with a length of 64. When the mmap() calls were\nchanged for the lengths to be based on the size of ExynosVideoMeta, the\ncorresponding munmap() calls were missed and left at 64.\n\n 1: https://github.com/LineageOS/android_hardware_samsung_slsi_exynos/blob/fb9857c5aa8923cc043b8f8a91bd24339c4bf528/libvppdisplay/ExynosDisplay.cpp#L1307\n\nChange-Id: Ib3591a8e952e0e1dbf1ccfc9aec36d908ca1db29","shortMessageHtmlLink":"libhwc1: Fix large mmap() and munmap() length mismatches"}},{"before":"0a9d736c23aa9f9287a392f133f489ccf1582999","after":"dbfd53607870de3000652ad0f0d930388dc50716","ref":"refs/heads/lineage-19.1","pushedAt":"2024-03-30T21:18:22.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: Fix large mmap() and munmap() length mismatches\n\nThe size of struct ExynosVideoMeta, used to determine lengths of\nmappings in mmap() calls, is very large compared to the length used in\ncorresponding munmap() calls - 64 bytes. This causes quite some memory\nleakage.\n\nUnder these BSP sources, ExynosVideoMeta has a size of 400 bytes. This\nmeans that, on every time mmap() and munmap() are called with their\nmismatched sizes in libhwc1, 400-64=336 bytes are undesirably leaked.\nWhile being very problematic on its own, in theory for real usage, that\ncould eventually cause problems particularly after many days of usage\nwith very heavy video playback (possibly even encountered by someone out\nthere, but unreported or ignored as a 'one-off thing'; no problems were\nknown before with these BSP sources).\n\nFix this problem by matching lengths used for the mmap() and munmap()\ncalls.\n\nAlso, don't map the entire size of the ExynosVideoMeta struct - Since\nonly the 'eType' member in it is checked in between the mmap() and\nmunmap() calls, and it's right at the start of the struct, the size of\nonly the 'eType' member can be mapped (and unmapped), which is only a\ntiny 4 bytes and seems to work fine.\n\nFor some trivia, this all looks like an oversight from someone at\nSamsung. In some really old BSP sources [1], mmap(), as well as\nmunmap(), is called with a length of 64. When the mmap() calls were\nchanged for the lengths to be based on the size of ExynosVideoMeta, the\ncorresponding munmap() calls were missed and left at 64.\n\n 1: https://github.com/LineageOS/android_hardware_samsung_slsi_exynos/blob/fb9857c5aa8923cc043b8f8a91bd24339c4bf528/libvppdisplay/ExynosDisplay.cpp#L1307\n\nChange-Id: Ib3591a8e952e0e1dbf1ccfc9aec36d908ca1db29","shortMessageHtmlLink":"libhwc1: Fix large mmap() and munmap() length mismatches"}},{"before":"ab998edcd0cde0778845b0086da6ae58b28527fe","after":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","ref":"refs/heads/lineage-21","pushedAt":"2024-03-30T15:02:26.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"ab998edcd0cde0778845b0086da6ae58b28527fe","after":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","ref":"refs/heads/lineage-20","pushedAt":"2024-03-30T15:02:20.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"ab998edcd0cde0778845b0086da6ae58b28527fe","after":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-03-30T15:02:14.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"ab998edcd0cde0778845b0086da6ae58b28527fe","after":"b1bee7fa0e9da5136d3de73224dfe2cb1f693bb7","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-03-30T15:02:08.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: use Strings8/16 c_str() and empty()\n\nAs per previous commits by David Marble <drmarble1@gmail.com> for\nAndroid 14 QPR2. (Though it's fine for 13 as well for keeping sources\nmore consistent.)\n\n $ grep -rli \"\\.string\" * | xargs -i@ sed -i 's/\\.string/\\.c_str/g' @\n $ grep -rli \"\\.isEmpty\" * | xargs -i@ sed -i 's/\\.isEmpty/\\.empty/g' @\n\nChange-Id: I2d4b8139a349e7bae676fead9018987e3aa8e602","shortMessageHtmlLink":"libhwc1: use Strings8/16 c_str() and empty()"}},{"before":"8850d590fb435c5f8e6d37efef969bec78105d5d","after":"ab998edcd0cde0778845b0086da6ae58b28527fe","ref":"refs/heads/lineage-21","pushedAt":"2024-03-30T12:24:12.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: Fix severe mmap() and munmap() length mismatches\n\nThe size of struct ExynosVideoMeta, used to determine lengths of\nmappings in mmap() calls, is very large compared to the length used in\ncorresponding munmap() calls - 64 bytes. This causes massive memory\nleakage.\n\nUnder the older BSP sources intended for Android 10, from where these\nHWC1 sources were imported here, ExynosVideoMeta had a size of 400\nbytes. Due to large additions to ExynosVideoMeta in these 13 sources,\nit has grown to 0x2438, or 9272(!), bytes, causing 9272-64=9208 bytes\nto be leaked every time mmap() and munmap() are called.\n\nBecause the leaking is so much more severe, it isn't so long in usage\nwith video playback before no more memory can be mapped, and that then\ncauses random soft reboots.\n\nFix this massive problem by matching lengths used for the mmap() and\nmunmap() calls.\n\nAlso, don't map the entire size of the ExynosVideoMeta struct - Since\nonly the 'eType' member in it is checked in between the mmap() and\nmunmap() calls, and it's right at the start of the struct, the size of\nonly the 'eType' member can be mapped (and unmapped), which is only a\ntiny 4 bytes and seems to work fine.\n\nFor some trivia, this all looks like an oversight from someone at\nSamsung. In some really old BSP sources [1], mmap(), as well as\nmunmap(), is called with a length of 64. When the mmap() calls were\nchanged for the lengths to be based on the size of ExynosVideoMeta, the\ncorresponding munmap() calls were missed and left at 64.\n\n 1: https://github.com/LineageOS/android_hardware_samsung_slsi_exynos/blob/fb9857c5aa8923cc043b8f8a91bd24339c4bf528/libvppdisplay/ExynosDisplay.cpp#L1307\n\nChange-Id: Ib3591a8e952e0e1dbf1ccfc9aec36d908ca1db29","shortMessageHtmlLink":"libhwc1: Fix severe mmap() and munmap() length mismatches"}},{"before":"8850d590fb435c5f8e6d37efef969bec78105d5d","after":"ab998edcd0cde0778845b0086da6ae58b28527fe","ref":"refs/heads/lineage-20","pushedAt":"2024-03-30T12:23:56.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: Fix severe mmap() and munmap() length mismatches\n\nThe size of struct ExynosVideoMeta, used to determine lengths of\nmappings in mmap() calls, is very large compared to the length used in\ncorresponding munmap() calls - 64 bytes. This causes massive memory\nleakage.\n\nUnder the older BSP sources intended for Android 10, from where these\nHWC1 sources were imported here, ExynosVideoMeta had a size of 400\nbytes. Due to large additions to ExynosVideoMeta in these 13 sources,\nit has grown to 0x2438, or 9272(!), bytes, causing 9272-64=9208 bytes\nto be leaked every time mmap() and munmap() are called.\n\nBecause the leaking is so much more severe, it isn't so long in usage\nwith video playback before no more memory can be mapped, and that then\ncauses random soft reboots.\n\nFix this massive problem by matching lengths used for the mmap() and\nmunmap() calls.\n\nAlso, don't map the entire size of the ExynosVideoMeta struct - Since\nonly the 'eType' member in it is checked in between the mmap() and\nmunmap() calls, and it's right at the start of the struct, the size of\nonly the 'eType' member can be mapped (and unmapped), which is only a\ntiny 4 bytes and seems to work fine.\n\nFor some trivia, this all looks like an oversight from someone at\nSamsung. In some really old BSP sources [1], mmap(), as well as\nmunmap(), is called with a length of 64. When the mmap() calls were\nchanged for the lengths to be based on the size of ExynosVideoMeta, the\ncorresponding munmap() calls were missed and left at 64.\n\n 1: https://github.com/LineageOS/android_hardware_samsung_slsi_exynos/blob/fb9857c5aa8923cc043b8f8a91bd24339c4bf528/libvppdisplay/ExynosDisplay.cpp#L1307\n\nChange-Id: Ib3591a8e952e0e1dbf1ccfc9aec36d908ca1db29","shortMessageHtmlLink":"libhwc1: Fix severe mmap() and munmap() length mismatches"}},{"before":"8850d590fb435c5f8e6d37efef969bec78105d5d","after":"ab998edcd0cde0778845b0086da6ae58b28527fe","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-03-30T12:23:47.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: Fix severe mmap() and munmap() length mismatches\n\nThe size of struct ExynosVideoMeta, used to determine lengths of\nmappings in mmap() calls, is very large compared to the length used in\ncorresponding munmap() calls - 64 bytes. This causes massive memory\nleakage.\n\nUnder the older BSP sources intended for Android 10, from where these\nHWC1 sources were imported here, ExynosVideoMeta had a size of 400\nbytes. Due to large additions to ExynosVideoMeta in these 13 sources,\nit has grown to 0x2438, or 9272(!), bytes, causing 9272-64=9208 bytes\nto be leaked every time mmap() and munmap() are called.\n\nBecause the leaking is so much more severe, it isn't so long in usage\nwith video playback before no more memory can be mapped, and that then\ncauses random soft reboots.\n\nFix this massive problem by matching lengths used for the mmap() and\nmunmap() calls.\n\nAlso, don't map the entire size of the ExynosVideoMeta struct - Since\nonly the 'eType' member in it is checked in between the mmap() and\nmunmap() calls, and it's right at the start of the struct, the size of\nonly the 'eType' member can be mapped (and unmapped), which is only a\ntiny 4 bytes and seems to work fine.\n\nFor some trivia, this all looks like an oversight from someone at\nSamsung. In some really old BSP sources [1], mmap(), as well as\nmunmap(), is called with a length of 64. When the mmap() calls were\nchanged for the lengths to be based on the size of ExynosVideoMeta, the\ncorresponding munmap() calls were missed and left at 64.\n\n 1: https://github.com/LineageOS/android_hardware_samsung_slsi_exynos/blob/fb9857c5aa8923cc043b8f8a91bd24339c4bf528/libvppdisplay/ExynosDisplay.cpp#L1307\n\nChange-Id: Ib3591a8e952e0e1dbf1ccfc9aec36d908ca1db29","shortMessageHtmlLink":"libhwc1: Fix severe mmap() and munmap() length mismatches"}},{"before":"8850d590fb435c5f8e6d37efef969bec78105d5d","after":"ab998edcd0cde0778845b0086da6ae58b28527fe","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-03-30T12:23:27.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"libhwc1: Fix severe mmap() and munmap() length mismatches\n\nThe size of struct ExynosVideoMeta, used to determine lengths of\nmappings in mmap() calls, is very large compared to the length used in\ncorresponding munmap() calls - 64 bytes. This causes massive memory\nleakage.\n\nUnder the older BSP sources intended for Android 10, from where these\nHWC1 sources were imported here, ExynosVideoMeta had a size of 400\nbytes. Due to large additions to ExynosVideoMeta in these 13 sources,\nit has grown to 0x2438, or 9272(!), bytes, causing 9272-64=9208 bytes\nto be leaked every time mmap() and munmap() are called.\n\nBecause the leaking is so much more severe, it isn't so long in usage\nwith video playback before no more memory can be mapped, and that then\ncauses random soft reboots.\n\nFix this massive problem by matching lengths used for the mmap() and\nmunmap() calls.\n\nAlso, don't map the entire size of the ExynosVideoMeta struct - Since\nonly the 'eType' member in it is checked in between the mmap() and\nmunmap() calls, and it's right at the start of the struct, the size of\nonly the 'eType' member can be mapped (and unmapped), which is only a\ntiny 4 bytes and seems to work fine.\n\nFor some trivia, this all looks like an oversight from someone at\nSamsung. In some really old BSP sources [1], mmap(), as well as\nmunmap(), is called with a length of 64. When the mmap() calls were\nchanged for the lengths to be based on the size of ExynosVideoMeta, the\ncorresponding munmap() calls were missed and left at 64.\n\n 1: https://github.com/LineageOS/android_hardware_samsung_slsi_exynos/blob/fb9857c5aa8923cc043b8f8a91bd24339c4bf528/libvppdisplay/ExynosDisplay.cpp#L1307\n\nChange-Id: Ib3591a8e952e0e1dbf1ccfc9aec36d908ca1db29","shortMessageHtmlLink":"libhwc1: Fix severe mmap() and munmap() length mismatches"}},{"before":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","after":"8850d590fb435c5f8e6d37efef969bec78105d5d","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-03-10T16:56:25.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","after":"8850d590fb435c5f8e6d37efef969bec78105d5d","ref":"refs/heads/lineage-20","pushedAt":"2024-03-10T16:56:22.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","after":"8850d590fb435c5f8e6d37efef969bec78105d5d","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-03-10T16:56:20.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","after":"8850d590fb435c5f8e6d37efef969bec78105d5d","ref":"refs/heads/lineage-21","pushedAt":"2024-03-10T16:56:18.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","after":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","ref":"refs/heads/lineage-20","pushedAt":"2024-02-24T10:17:38.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","after":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-02-24T10:17:30.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","after":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-02-24T10:17:23.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","after":"7ef320ffbbe241c19a689026bab27b4d43d02ac0","ref":"refs/heads/lineage-21","pushedAt":"2024-02-24T10:17:14.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"2cf8aaee2485bf9657b835edafff97581121a40f","after":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","ref":"refs/heads/lineage-21","pushedAt":"2024-01-31T18:20:50.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"2cf8aaee2485bf9657b835edafff97581121a40f","after":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","ref":"refs/heads/lineage-20","pushedAt":"2024-01-31T18:20:45.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"2cf8aaee2485bf9657b835edafff97581121a40f","after":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-01-31T18:20:39.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"2cf8aaee2485bf9657b835edafff97581121a40f","after":"dedbe77436b1ffe52f9508663a089b6158a9d4eb","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-01-31T18:20:32.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"4e67f6346c39a7296c0269b0f0313cea7c619013","after":"2cf8aaee2485bf9657b835edafff97581121a40f","ref":"refs/heads/lineage-20_13-e850-96","pushedAt":"2024-01-24T21:19:01.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}},{"before":"4e67f6346c39a7296c0269b0f0313cea7c619013","after":"2cf8aaee2485bf9657b835edafff97581121a40f","ref":"refs/heads/lineage-21_13-e850-96","pushedAt":"2024-01-24T21:18:53.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"K9100ii","name":"K9100ii","path":"/K9100ii","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/55348214?s=80&v=4"},"commit":{"message":"graphics: Add HWC1 from Android 10 BSP sources\n\nspecifically my sources at commit 8e8a2e09f9 (\"libhwc1: libhwcService:\nifdef out two missed virtual display functions\"):\nhttps://github.com/K9100ii/android_hardware_samsung_slsi-linaro_graphics/tree/8e8a2e09f9463bb0498c8c5ac756626f449e4fd2\n\nMore specifically, while this mainly pulls in libhwc1 from those\nsources, with all the fixes I applied to it, this also:\n - pulls libmpp in, which HWC1 depends on, and sets the flag for using\n   legacy lib[g]scaler as long as HWC1 is used as is necessary,\n - adds an include in libion, without which there are compilation errors,\n - adds the exynos7870 and exynos8890 SoC-specific HWC\n   configurations,\n - and finally, puts the inclusion of the libhwjpeg here behind the\n   TARGET_USES_UNIVERSAL_LIBHWJPEG flag, for having different custom\n   copies behind different flags in device trees.\nSome code within libhwc1 and the SoC-specific configs\nthat depended on extra definitions in a custom hwcomposer_defs.h header\nhas also been removed, and having it all in vendor is now unconditional.\n\nCompared to my Android 10 BSP sources, Exynos 7580 support has been left\nout, simply since I haven't looked at keeping its OSS libMcRegistry\naround alongside the other newer versions, as it was kept in the Android\n10 sources, and I doubt I'll be doing any testing for it. Nevertheless,\nit would be possible to adapt my changes there to these sources for\nanyone with any interest.\n\nThis has all only been tested on exynos7870, while exynos8890\nconfiguration has been added as well since it should just work the same\nand there are no extras like the libMcRegistry on 7580.\n\nThis makes these sources a drop-in replacement for the Android 10 ones,\nasides from the 7580 support.\n\nChange-Id: Ia5f12e72a003b513717b01519d43648485414aa3","shortMessageHtmlLink":"graphics: Add HWC1 from Android 10 BSP sources"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"startCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wNi0yNVQwODozNDozOC4wMDAwMDBazwAAAARuXUs8","endCursor":"Y3Vyc29yOnYyOpK7MjAyNC0wMS0yNFQyMToxODo1My4wMDAwMDBazwAAAAPoVhZk"}},"title":"Activity · K9100ii/android_hardware_samsung_slsi-linaro_graphics"}