I'm a bit behind on the devel kernels. I'm currently running 2906 (2.6.20rc4)...
In my BZ (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=219900) I found (and got a fix for) the case were doing a "cp -a src dest" resulted in the dest file being zero length.
Has this been apply to the most recent kernels (either devel or vanilla)? Should I bring this up on LKML for 2.6.20?
-Paul
On Sat, 27 Jan 2007 14:31:29 -0700, Paul Dickson wrote:
I'm a bit behind on the devel kernels. I'm currently running 2906 (2.6.20rc4)...
In my BZ (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=219900) I found (and got a fix for) the case were doing a "cp -a src dest" resulted in the dest file being zero length.
Has this been apply to the most recent kernels (either devel or vanilla)? Should I bring this up on LKML for 2.6.20?
The patch adds NFS support for SELinux xattrs.
-Paul
On Sat, 2007-01-27 at 14:54 -0700, Paul Dickson wrote:
On Sat, 27 Jan 2007 14:31:29 -0700, Paul Dickson wrote:
I'm a bit behind on the devel kernels. I'm currently running 2906 (2.6.20rc4)...
In my BZ (https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=219900) I found (and got a fix for) the case were doing a "cp -a src dest" resulted in the dest file being zero length.
Has this been apply to the most recent kernels (either devel or vanilla)? Should I bring this up on LKML for 2.6.20?
The patch adds NFS support for SELinux xattrs.
No, it makes NFS lie about setting the SELinux attribute, while not truly setting it at all (other than in-core). So it is then lost when the inode is evicted from memory or the system reboots. The patch was rejected upstream. The problem lies in cp (coreutils), not NFS, IIUC - it is making faulty assumptions about setxattr() always being supported when getxattr() is.