Hello.
I still object to lying to the program by returning success instead of
failure, EPERM, or something like that. It at least gives the program an
opportunity to fail gracefully if it was not actually malicious. Lying
to the program and returning success is suitable if you are trying to
On Wed, 2007-09-12 at 17:51 +0900, Yuichi Nakamura wrote:
Hi.
Stephen Smalley pointed out possibility of race condition
in off-list discussion.
Stephen Smalley said:
One other observation about the patch: it presently leaves open a
(small) race window in which the file could get
On Thu, 13 Sep 2007 08:58:32 -0400
Stephen Smalley wrote:
On Wed, 2007-09-12 at 17:51 +0900, Yuichi Nakamura wrote:
snip
Thanks, a few comments below.
Thanks for comments!
* Description of patch
This patch improves performance of read/write in SELinux.
It improves performance by
Hello.
I would like to propose patch that reduces overhead in read/write by SELinux.
I sent RFC in previous thread.
http://lkml.org/lkml/2007/9/6/14
As a result of discussion in previous thread,
quality of code has improved, so I would like to submit patch here.
1. Background
Look at benchmark
