Comment 31 for bug 242448

I have reproduced the original problem using the 2.6.24-19.41 generic kernel image from the hardy-updates repository, and can confirm that the 2.6.24-21.42 generic kernel image from hardy-proposed does eliminate the file corrupting behavior, at least when using ecryptfs with the passphrase option.

I also exercised the ecryptfs filesystem by running LTP's filesystem stress tests manually on the encryptfs filesystem and saw no regressions, for what it's worth. (Admittedly, the LTP tests didn't catch this bug.)

Serge, the reason Leann's comment was added to this bug is because the development (intrepid) task's state in launchpad is currently set to "Confirmed" or in other words is still considered an open bug against the development kernel. If you or anyone else can confirm that it was indeed fixed upstream before 2.6.26, then please mark that task as invalid. Thanks!