Although linux allows disk addressing in 512-byte chunks, it only tracks
device SIZE in 1024-byte increments.
Thus, if we try to access the last, odd, 512-byte block on a partition,
linux will tell us we're past the end of the disk even if we're not,
since it rounds down to the nearest multiple of 1024 bytes.
This will only matter when: multiple blocksizes get implemented,
512-byte blocks are selected, the filesystem fills the partition,
and the partition size is not a 1024-byte multiple.
This is not currently a problem, since mkfs rounds the size down to a
multiple of the filesystem blocksize, which is currently 4096 on
x86. This results in a nice 1024 multiple as well.
So this is just planning ahead:
force even number of 512-byte blocks when data/log/rt fills partition
Date: Thu Jan 3 13:39:35 PST 2002
Workarea: stout.americas.sgi.com:/localhome/eric/2.4.x-xfs/workarea-reallyclean
The following file(s) were checked into:
bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs
Modid: xfs-cmds:slinx:109087a
cmd/xfsprogs/mkfs/xfs_mkfs.c - 1.19
cmd/xfsprogs/growfs/xfs_growfs.c - 1.8
|