On Thu, May 01, 2025 at 12:52:08PM -0700, Darrick J. Wong wrote: > From: Darrick J. Wong <djwong@xxxxxxxxxx> > > It's silly to call xfs_setsize_buftarg from xfs_alloc_buftarg with the > block device LBA size because we don't need to ask the block layer to > validate a geometry number that it provided us. Instead, set the > preliminary bt_meta_sector* fields to the LBA size in preparation for > reading the primary super. > > It's ok to lose the sync_blockdev call at buftarg creation time for the > external log and rt devices because we don't read from them until after > calling xfs_setup_devices. We do need an explicit sync for the data > device because we read the primary super before calling > xfs_setup_devices. Should we just it for all of them in open_devices now that the sync is decoupled from setting the block size? Otherwise this looks good, but I guess this should go before the atomic writes series in the end?
