J R, I did get around to trying the patch. It did not patch cleanly against aufs2/34. The patch looks for "!!" in various places where as the original doesn't contain these "return !!au_test_subdir(h_d1, h_d2)"?
Not sure what that's supposed to signify, but I went ahead an applied the changes by hand in branch.c and loop.c. My own testing concurs with your results. The "is overlapped" restrictions have been removed for mount points. I deployed this version onto a system which runs aufs as root, and this version worked without a hitch. I also tested cases which should fail, such as a real directory overlap. > mkdir -p a/b > mount -t aufs aufs -o br:a:a/b # is overlap error Even tried to trick aufs. > mkdir -p a/b b > mount --bind a/b b > mount -t aufs aufs union -o br:a:b # is overlap error It looks good to me. On 07/20/2010 10:12 PM, sf...@users.sourceforge.net wrote: > sf...@users.sourceforge.net: > >> Lou Gosselin: >> >>> Test Case E # This won't work without the modified aufs... >>> > mkdir -p union a/mnt/original_files >>> > mount -t tmpfs tmpfs a/mnt >>> > mkdir -p a/mnt/new_mount_files >>> > mount -t aufs aufs union -o br:a:a/mnt >>> > find >>> >> I am reviewing the behaviour of overlapping, and found case E works too. >> Which version of aufs (and kernel) are you using? >> >> $ mkdir -p u a/mnt >> $ sudo mount -t tmpfs nont a/mnt >> $ sudo mount -t aufs -o br:a:a/mnt none u >> $ >> > Lou and Barry, > > Here is a new specification and a patch about overlapping, please review > and test it if you have time. > > - /au = /br0 + /br1 ... + /brN > - /brX/dirA exists > - /brX/fs.img exists > > old-spec new-spec > -------------------------------------------------- > append:/out/dirB allow allow > append:/brX/dirA x x (Lou's case C) > append:/au/dirA x x > > mount something at /brX/dirA, > append:/brX/dirA allow allow (Lou's case E?) > > mount something at /au/dirA, > append:/au/dirA allow allow > > loopback mount /brX/fs.img at /out/dirB, > append:/out/dirB x ALLOW (Barry may be interested) > > loopback mount /au/fs.img at /out/dirB, > append:/out/dirB x x > > > J. R. Okajima > > ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first