On Thu, Mar 27, 2008 at 07:13:50AM +0100, Andi Kleen wrote: > Bob Copeland <me@xxxxxxxxxxxxxxx> writes: > > + > > + bitmap_size = (sbi->s_num_blocks + 7) / 8; > > + array_size = (bitmap_size + sb->s_blocksize - 1) / sb->s_blocksize; > > + > > + sbi->s_imap_size = array_size; > > + sbi->s_imap = kzalloc(array_size * sizeof(unsigned long), GFP_KERNEL); > > > If the array can be really 300k you should probably have a vmalloc fallback > here. bitmap_size will be 300k (20G / 8192 / 8), but array_size adds another level of indirection so s_imap will be 300k/8192 segments, each of size 8192. Hmm, I guess that should be 'sizeof(unsigned long *)' above. One might wonder why we keep it in memory anyway since we can read the bitmap blocks on demand. The reason is that once the tree walk is added for ReplayTV then it would use this same i_map. However, I've yet to hear from interested RTV folks so it is currently read-only and there is no tree walk for the bitmap. -- Bob Copeland %% www.bobcopeland.com -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
