On Thu, Jan 09, 2020 at 07:40:41AM +1100, Dave Chinner wrote:
> On Tue, Jan 07, 2020 at 08:17:38PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> >
> > Introduce a new #define for the maximum supported file block offset.
> > We'll use this in the next patch to make it more obvious that we're
> > doing some operation for all possible inode fork mappings after a given
> > offset. We can't use ULLONG_MAX here because bunmapi uses that to
> > detect when it's done.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> > ---
> > fs/xfs/libxfs/xfs_format.h | 1 +
> > fs/xfs/xfs_reflink.c | 3 ++-
> > 2 files changed, 3 insertions(+), 1 deletion(-)
> >
> >
> > diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
> > index 1b7dcbae051c..c2976e441d43 100644
> > --- a/fs/xfs/libxfs/xfs_format.h
> > +++ b/fs/xfs/libxfs/xfs_format.h
> > @@ -1540,6 +1540,7 @@ typedef struct xfs_bmdr_block {
> > #define BMBT_BLOCKCOUNT_BITLEN 21
> >
> > #define BMBT_STARTOFF_MASK ((1ULL << BMBT_STARTOFF_BITLEN) - 1)
> > +#define XFS_MAX_FILEOFF (BMBT_STARTOFF_MASK)
>
> Isn't the maximum file offset in the BMBT the max start offset + the
> max length of the extent that is located at BMBT_STARTOFF_MASK?
Apologies for responding to a question with another question, but has
there ever been an XFS that supported an inode size of more than 8EB?
Linux supports at most a file offset of 8EB, which is 2^63-1, or
0x7FFF,FFFF,FFFF,FFFF. On a filesystem with 512-byte blocks, the very
last byte in the file would be in block 2^54-1, or 0x3F,FFFF,FFFF,FFFF.
Larger blocksizes decrease that even further (e.g. 2^47-1, or
0x7FFF,FFFF,FFFF on 64k block filesystems).
Therefore, on Linux I conclude that the largest file offset (block)
possible is 2^54-1, which is BMBT_STARTOFF_MASK. Unless there's an
XFS port that actually supports 16EB files, BMBT_STARTOFF_MASK will
suffice here.
--D
> Cheers,
>
> Dave.
> --
> Dave Chinner
> david@xxxxxxxxxxxxx
