From de994667dda925109e862edadb4aa4feaecd0e6b Mon Sep 17 00:00:00 2001 From: Alex Zhuravlev Date: Thu, 28 Nov 2019 15:04:25 +0300 Subject: [PATCH] LU-12988 ldiskfs: skip non-loaded groups at cr=0/1 cr=0 is supposed to be an optimization to save CPU cycles, but if buddy data (in memory) is not initialized then all this makes no sense as we have to do sync IO taking a lot of cycles. also, at cr=0 mballoc doesn't store any avaibale chunk. cr=1 also skips groups using heruistic based on avg. fragment size. it's more useful to skip such groups and switch to cr=2 where groups will be scanned for available chunks. using sparse image and dm-slow virtual device of 120TB was simulated. then the image was formatted as OST and filled using debugfs to mark ~85% of available space as busy. mount as OST w/o the patch couldn't complete in half an hour (according to vmstat it would take ~10-11 hours). with the patch applied mount took ~20 seconds. Lustre-change: https://review.whamcloud.com/36891 Lustre-commit: 6a7a700a1490dfde6b60c2fb36df92a052059866 Signed-off-by: Alex Zhuravlev Change-Id: I88c8c1b01b386af0fa438bfeb97acb6110bd00ec Reviewed-by: Andreas Dilger Reviewed-by: Wang Shilong Reviewed-by: Artem Blagodarenko Signed-off-by: Minh Diep Reviewed-on: https://review.whamcloud.com/37539 Tested-by: jenkins Tested-by: Maloo Reviewed-by: Oleg Drokin --- .../rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch | 19 +++++++++++++++++++ .../rhel8/ext4-mballoc-skip-uninit-groups-cr0.patch | 19 +++++++++++++++++++ .../kernel_patches/series/ldiskfs-3.10-rhel7.2.series | 1 + .../kernel_patches/series/ldiskfs-3.10-rhel7.3.series | 1 + .../kernel_patches/series/ldiskfs-3.10-rhel7.4.series | 1 + .../kernel_patches/series/ldiskfs-3.10-rhel7.5.series | 1 + .../kernel_patches/series/ldiskfs-3.10-rhel7.6.series | 1 + .../kernel_patches/series/ldiskfs-3.10-rhel7.7.series | 1 + .../kernel_patches/series/ldiskfs-3.10-rhel7.series | 1 + 9 files changed, 45 insertions(+) create mode 100644 ldiskfs/kernel_patches/patches/rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch create mode 100644 ldiskfs/kernel_patches/patches/rhel8/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/patches/rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch b/ldiskfs/kernel_patches/patches/rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch new file mode 100644 index 0000000..1bb8398 --- /dev/null +++ b/ldiskfs/kernel_patches/patches/rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch @@ -0,0 +1,19 @@ +--- linux-4.18/fs/ext4/mballoc.c 2019-11-28 14:55:26.500545920 +0300 ++++ linux-4.18/fs/ext4/mballoc.c 2019-11-28 14:53:18.600086008 +0300 +@@ -2060,7 +2060,15 @@ static int ext4_mb_good_group(struct + + /* We only do this if the grp has never been initialized */ + if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { +- int ret = ext4_mb_init_group(ac->ac_sb, group); ++ int ret; ++ ++ /* cr=0/1 is a very optimistic search to find large ++ * good chunks almost for free. if buddy data is ++ * not ready, then this optimization makes no sense */ ++ ++ if (cr < 2) ++ return 0; ++ ret = ext4_mb_init_group(ac->ac_sb, group); + if (ret) + return ret; + } diff --git a/ldiskfs/kernel_patches/patches/rhel8/ext4-mballoc-skip-uninit-groups-cr0.patch b/ldiskfs/kernel_patches/patches/rhel8/ext4-mballoc-skip-uninit-groups-cr0.patch new file mode 100644 index 0000000..38925ac --- /dev/null +++ b/ldiskfs/kernel_patches/patches/rhel8/ext4-mballoc-skip-uninit-groups-cr0.patch @@ -0,0 +1,19 @@ +--- linux-4.18/fs/ext4/mballoc.c 2019-11-28 14:55:26.500545920 +0300 ++++ linux-4.18/fs/ext4/mballoc.c 2019-11-28 14:53:18.600086008 +0300 +@@ -2060,7 +2060,15 @@ static int ext4_mb_good_group(struct + + /* We only do this if the grp has never been initialized */ + if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) { +- int ret = ext4_mb_init_group(ac->ac_sb, group, GFP_NOFS); ++ int ret; ++ ++ /* cr=0/1 is a very optimistic search to find large ++ * good chunks almost for free. if buddy data is ++ * not ready, then this optimization makes no sense */ ++ ++ if (cr < 2) ++ return 0; ++ ret = ext4_mb_init_group(ac->ac_sb, group, GFP_NOFS); + if (ret) + return ret; + } diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.2.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.2.series index ac83b2f..dfbd2e4 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.2.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.2.series @@ -38,3 +38,4 @@ rhel7/ext4-export-orphan-add.patch rhel7/ext4-mmp-dont-mark-bh-dirty.patch rhel7/ext4-include-terminating-u32-in-size-of-xattr-entries-when-expanding-inodes.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.3.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.3.series index 3dfecb7..3b04032 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.3.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.3.series @@ -38,3 +38,4 @@ rhel7/ext4-export-orphan-add.patch rhel7/ext4-mmp-dont-mark-bh-dirty.patch rhel7/ext4-include-terminating-u32-in-size-of-xattr-entries-when-expanding-inodes.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.4.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.4.series index 66035ef..fd3e90d 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.4.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.4.series @@ -38,3 +38,4 @@ rhel7/ext4-export-orphan-add.patch rhel7/ext4-mmp-dont-mark-bh-dirty.patch rhel7/ext4-include-terminating-u32-in-size-of-xattr-entries-when-expanding-inodes.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.5.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.5.series index 6c4e7dd..84307c9 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.5.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.5.series @@ -38,3 +38,4 @@ rhel7/ext4-mmp-dont-mark-bh-dirty.patch rhel7/ext4-include-terminating-u32-in-size-of-xattr-entries-when-expanding-inodes.patch rhel7.2/ext4-export-mb-stream-allocator-variables.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.6.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.6.series index d10b17f..94a4c07 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.6.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.6.series @@ -39,3 +39,4 @@ rhel7/ext4-include-terminating-u32-in-size-of-xattr-entries-when-expanding-inode rhel7.2/ext4-export-mb-stream-allocator-variables.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch rhel7.2/ext4-simple-blockalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.7.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.7.series index ca67087..47592d7 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.7.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.7.series @@ -39,3 +39,4 @@ rhel7.2/ext4-export-mb-stream-allocator-variables.patch rhel7.7/ext4-fix-project-with-unpatched-kernel.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch rhel7.2/ext4-simple-blockalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch diff --git a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.series b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.series index 2f67631..52228ba 100644 --- a/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.series +++ b/ldiskfs/kernel_patches/series/ldiskfs-3.10-rhel7.series @@ -33,3 +33,4 @@ rhel7/ext4-export-orphan-add.patch rhel7/ext4-mmp-dont-mark-bh-dirty.patch rhel7/ext4-include-terminating-u32-in-size-of-xattr-entries-when-expanding-inodes.patch rhel7/ext4-optimize-ext4_find_delalloc_range-in-nodelalloc.patch +rhel7/ext4-mballoc-skip-uninit-groups-cr0.patch -- 1.8.3.1