On 08/27/2009 06:54 AM, Eric Blake wrote:
I hope that my example shows why doing it in the kernel is desirable - there is no safe way to keep the pre-O_CLOEXEC efficiency using just the library, but there IS a way to do it with kernel support:
You're describing a very special case where the performance implications are really minimal and try to argue that is a good enough reason? I don't think so.
If a program really has to do thousands of these safe open calls then it can invest time into opening /dev/null for any of the unallocated descriptors < 3. You can even embed this logic in the safer_open function.
-- ➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
