Hi all, now, that was easy. Thanks to willy and his recent patchset to support large folios in gfs2 turns out that most of the work to support high-order folios for I/O is actually done. It only need twe rather obvious patches to allocate folios with the order derived from the mapping blocksize, and to adjust readahead to avoid reading off the end of the device. But with these two patches (and the patchset from hch to switch the block device over to iomap) (and the patchset from ritesh to support sub-blocksize iomap buffers) I can now do: # modprobe brd rd_size=524288 rd_blksize=16384 # mkfs.xfs -b size=16384 /dev/ram0 it still fails when trying to mount the device: XFS (ram0): Cannot set_blocksize to 16384 on device ram0 but to my understanding this is being worked on. Christoph, any chance to have an updated submission of your patchset to convert block devices over to iomap? I don't actually need the last one to switch off buffer heads, but the others really do help for this case. The entire tree can be found at: git.kernel.org:/pub/scm/linux/git/kernel/hare/scsi-devel.git branch brd.v2 Happy hacking! Hannes Reinecke (6): brd: convert to folios brd: abstract page_size conventions brd: make sector size configurable brd: make logical sector size configurable mm/filemap: allocate folios with mapping blocksize mm/readahead: align readahead down to mapping blocksize Pankaj Raghav (1): brd: use XArray instead of radix-tree to index backing pages drivers/block/brd.c | 320 +++++++++++++++++++++------------------- include/linux/pagemap.h | 7 + mm/filemap.c | 7 +- mm/readahead.c | 10 +- 4 files changed, 186 insertions(+), 158 deletions(-) -- 2.35.3
