Re: Squeeze on USB
Ben Hutchings:
> I was hoping you would be able to identify the most important fixes,
> dealing with data loss and security bugs. This long list of changes
> appears to include cleanup and new features, which would not be
No, such changes are removed.
You may be able to see what I have removed if you run
$ cd aufs2-2.6.git
$ git log --no-merges --since=2010/01/24 aufs2.1-32
$ git log --no-merges aufs2.1-32..aufs2.2-35 fs/aufs
Most of the list will surely cause deadlock or crash in certain
circumstances.
You may want to ask "what is certain circumstance? how often will users
meet such problem?"
When I write "possible bugfix", it means there are some conditions to
reproduce the bug, for instance,
- enable CONFIG_AUFS_xxx.
- run processA which issues systemcallA.
- run processB which issues systemcallB.
- when aufs operates systemcallA and systemcallB at the same time, a
deadlock will happen.
In other words, if you disable CONFIG_AUFS_xxx, or issue both of
systemcallA/B but their timing is different, then you will never meet
the problem. It is totally up to your configuration and operation (or
others).
For security fixes, it is also up to you.
For example,
dcb6ad5 2011-03-01 Fix aufs call of security_path_mknod
Author: John Johansen <john.johansen@canonical.com>
is to fix an incorrect parameter for security_path_mknod().
As you know, security_path_mknod() has its meaning only when
- you configure kernel to do so.
- and users sets something to do so.
Again, if one of these conditions are not true, security_path_mknod()
becomes empty and the commit has no meaning.
Finally, I'd suggest you to
- make aufs in debian as the last aufs2.1-32 unconditionally.
- and choose the commits in aufs2.2 (the last part of my previous mail).
J. R. Okajima
Reply to: