On Fri, Nov 21, 2008 at 11:11:10PM +0530, Aneesh Kumar K.V wrote: > On Fri, Nov 21, 2008 at 11:29:49AM -0600, Eric Sandeen wrote: > > Aneesh Kumar K.V wrote: > > > Rename the lower bits with suffix _lo and add helper > > > to access the values. Also rename bg_itable_unused_hi > > > to bg_pad as in e2fsprogs. bg_itable_unused_hi is never > > > used before > > > > Maybe you can add info about under which circumstances (what sort of > > fielsystem geometry) the hi bits are needed? And the failure mode when > > we don't use them, if that's known? > > Coly actually said it was added for inode reservation. But he agreed to > drop the high bits now and add them later once the performance issue > with the patch is addressed. The bg_itable_unused field is used for the lazy initialization of inodes. If the units of this field was blocks was blocks, we wouldn't need bg_itable_unused_hi for quite a larger set of blocks. I don't think we should nuke bg_itable_unused_hi, especially given that we have a patch to support the kernel initializing the inodes (yet to be reviewed). As near as I can tell at the time when the fields were added to e2fsprogs, we hadn't yet assigned the pad field to bg_itable_unused_hi, but it's clearly needed if we want to support a larger number of blocks or inodes per block group. Instead we should create the helper functions for bg_itable_unused, and update e2fsprogs to also support these fields. Adding e2fsprogs for large block groups is lower priority than large block numbers, though. - Ted -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
