For the default filesystem configuration anyway. There was one remaining
path in the default block sizes allowed on linux where we need to use
address space manipulation. This removes that case and turns off the
remapping code by default. Dealing with filesystems with a block
size bigger than a page will still require something like this
feature, and dealing with directory blocks bigger than a page will
as well (this is turned off in mkfs at the moment).
Date: Mon Mar 12 11:22:23 PST 2001
Workarea: jen.americas.sgi.com:/src/lord/xfs-linux.2.4
The following file(s) were checked into:
bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs
Modid: 2.4.x-xfs:slinx:89552a
linux/fs/xfs/xfs_rw.c - 1.334
- allow better flag passing down from xfs to pagebuf
linux/fs/xfs/xfs_buf.h - 1.68
- define xfs_biomove to call pagebuf_iomove, remove definition of
xfs_bp_mapin
linux/fs/xfs/xfs_attr.c - 1.82
- Use xfs_biomove to move extended attribute data in and out of buffers.
linux/include/linux/page_buf.h - 1.81
- define pagebuf_iomove, remove mapin and mapout prototypes
linux/fs/pagebuf/page_buf.c - 1.64
- Add pagebuf_iomove to copy arbiraty sized chunks of memory in and
out of pagebufs without mapping the memory. Make all the code to
do with address space remapping not compile by default.
|