Hi all, This patchset mainly provides a new helper z_erofs_get_max_pclustersize() to prepare for per-file compression strategies, valid pclustersize can be returned according to detailed data type or access patterns.
In order to do that, compression header is now generated on the per-file basis as well, which will be also needed for parallel compression in the future. Note that https://lore.kernel.org/r/[email protected] should be applied after "erofs-utils: compress trailing data for big pcluster properly" is used or some compress indexes won't be parsed correctly. Thanks, Gao Xiang Gao Xiang (4): erofs-utils: compress trailing data for big pcluster properly erofs-utils: reserve physical_clusterbits[] erofs-utils: prepare for per-(sub)file compress strategies erofs-utils: sync up z_erofs_get_extent_compressedlen() include/erofs/internal.h | 1 - lib/compress.c | 91 +++++++++++++++++++++++++++------------- lib/compressor.c | 5 ++- lib/zmap.c | 53 +++++++++++++---------- 4 files changed, 96 insertions(+), 54 deletions(-) -- 2.20.1
