I was chasing a corruption problem which happens on small
block filesystems. This fixes an oops I found along the
way, and reduces the frequency of the corruption. The
real problem is still out there though....
Date: Wed Aug 21 13:29:14 PDT 2002
The following file(s) were checked into:
linux/fs/xfs/linux/xfs_iops.c - 1.171
- set BH_New in less cases
linux/fs/xfs/pagebuf/page_buf_io.c - 1.54
- make the release_page path aware of end of file and remove delalloc
on buffer heads in this case.
linux/fs/xfs/pagebuf/page_buf.h - 1.39
- pagebuf_release_page returns a value now