>> Modified: sys/ufs/ffs fs.h
>> Fixed lblktosize(). It overflowed at 2G. This bug only affected
>> ufs_read() and ufs_write().
> Are you going to bring this into -stable, or do you think it is unnecessary
>since we don't support >2GB files there anyway due to VM page restrictions?
I didn't plan to. Perhaps we should set fs_maxfilesize to 2G-1. This
should be safer than doing nothing or attempting to trap all the overflows.
I thought that large files worked in -current. A long time ago, I put a
1 TB file system on /dev/vn0. The metadata must have gone beyond 2G
logical in the underlying file.