From: Pankaj Raghav <p.raghav@xxxxxxxxxxx> This is the 13th version of the series that enables block size > page size (Large Block Size) experimental support in XFS. Please consider this for the inclusion in 6.12. The context and motivation can be seen in cover letter of the RFC v1 [0]. We also recorded a talk about this effort at LPC [1], if someone would like more context on this effort. Thanks to David Howells, the page cache changes have also been tested on top of AFS[2] with mapping_min_order set. A lot of emphasis has been put on testing using kdevops, starting with an XFS baseline [3]. The testing has been split into regression and progression. Regression testing: In regression testing, we ran the whole test suite to check for regressions on existing profiles due to the page cache changes. I also ran split_huge_page_test selftest on XFS filesystem to check for huge page splits in min order chunks is done correctly. No regressions were found with these patches added on top. Progression testing: For progression testing, we tested for 8k, 16k, 32k and 64k block sizes. To compare it with existing support, an ARM VM with 64k base page system (without our patches) was used as a reference to check for actual failures due to LBS support in a 4k base page size system. No new failures were found with the LBS support. We've done some preliminary performance tests with fio on XFS on 4k block size against pmem and NVMe with buffered IO and Direct IO on vanilla Vs + these patches applied, and detected no regressions. We ran sysbench on postgres and mysql for several hours on LBS XFS without any issues. We also wrote an eBPF tool called blkalgn [5] to see if IO sent to the device is aligned and at least filesystem block size in length. For those who want this in a git tree we have this up on a kdevops large-block-minorder-for-next-v13 tag [6]. [0] https://lore.kernel.org/lkml/20230915183848.1018717-1-kernel@xxxxxxxxxxxxxxxx/ [1] https://www.youtube.com/watch?v=ar72r5Xf7x4 [2] https://lore.kernel.org/linux-mm/3792765.1724196264@xxxxxxxxxxxxxxxxxxxxxx/ [3] https://github.com/linux-kdevops/kdevops/blob/master/docs/xfs-bugs.md 489 non-critical issues and 55 critical issues. We've determined and reported that the 55 critical issues have all fall into 5 common XFS asserts or hung tasks and 2 memory management asserts. [4] https://github.com/linux-kdevops/fstests/tree/lbs-fixes [5] https://github.com/iovisor/bcc/pull/4813 [6] https://github.com/linux-kdevops/linux/ [7] https://lore.kernel.org/linux-kernel/Zl20pc-YlIWCSy6Z@xxxxxxxxxxxxxxxxxxxx/#t Changes since v12: - Fixed the issue of masking the wrong bits in PATCH 1. - Collected Tested-by from David Howells. Dave Chinner (1): xfs: use kvmalloc for xattr buffers Luis Chamberlain (1): mm: split a folio in minimum folio order chunks Matthew Wilcox (Oracle) (1): fs: Allow fine-grained control of folio sizes Pankaj Raghav (7): filemap: allocate mapping_min_order folios in the page cache readahead: allocate folios with mapping_min_order in readahead filemap: cap PTE range to be created to allowed zero fill in folio_map_range() iomap: fix iomap_dio_zero() for fs bs > system page size xfs: expose block size in stat xfs: make the calculation generic in xfs_sb_validate_fsb_count() xfs: enable block size larger than page size support fs/iomap/buffered-io.c | 4 +- fs/iomap/direct-io.c | 45 +++++++++++-- fs/xfs/libxfs/xfs_attr_leaf.c | 15 ++--- fs/xfs/libxfs/xfs_ialloc.c | 5 ++ fs/xfs/libxfs/xfs_shared.h | 3 + fs/xfs/xfs_icache.c | 6 +- fs/xfs/xfs_iops.c | 2 +- fs/xfs/xfs_mount.c | 8 ++- fs/xfs/xfs_super.c | 28 +++++--- include/linux/huge_mm.h | 14 ++-- include/linux/pagemap.h | 123 ++++++++++++++++++++++++++++++---- mm/filemap.c | 36 ++++++---- mm/huge_memory.c | 60 +++++++++++++++-- mm/readahead.c | 83 +++++++++++++++++------ 14 files changed, 347 insertions(+), 85 deletions(-) base-commit: eb8c5ca373cbb018a84eb4db25c863302c9b6314 -- 2.44.1
