[PATCH 0/3] updated *at function patch

[PATCH 0/3] updated *at function patch

[Ulrich Drepper]

[Sorry about the previous mails, I've bad problems with my connection
and need to resort to doing this remotely with very high latency.]

Just a little update and relative to the current sources.

The only code change are three lines of code: renaming fchmod to fchmodat,
adding a new parameter, passing it to __user_walk_fd, and a new one-line
wrapper function sys_chmod.  The other changes are hooking up the syscall
and the previously not-hooked-up faccessat syscal.

Re: [PATCH 0/3] updated *at function patch

[Andrew Morton]

I renumbered the syscalls due to the queued-up sys_migrate_pages.

We don't have changelogs for these patches.  Apart from the usual
what-it-does and how-it-does it I'd really like to be reminded of the
"why".  Adding a bunch of stuff to the core kernel codepaths needs to have
a good reason and I've forgotten your rationale.

We'll need Signed-off_by:'s for all these patches.

In future, please avoid sending multiple patches under the same Subject:,
thanks.

Re: [PATCH 0/3] updated *at function patch

[Ulrich Drepper]

[-- Attachment #1: Type: text/plain, Size: 1912 bytes --]

Andrew Morton wrote:
> We don't have changelogs for these patches.  Apart from the usual
> what-it-does and how-it-does it I'd really like to be reminded of the
> "why".  Adding a bunch of stuff to the core kernel codepaths needs to have
> a good reason and I've forgotten your rationale.

This is the intro from the first set of patches I sent (I adjusted it
for the additional syscalls):

====
Here is a series of patches which introduce in total 13 new system calls
which take a file descriptor/filename pair instead of a single file
name.  These functions, openat etc, have been discussed on numerous
occasions.  They are needed to implement race-free filesystem traversal,
they are necessary to implement a virtual per-thread current working
directory (think multi-threaded backup software), etc.

We have in glibc today implementations of the interfaces which use the
/proc/self/fd magic.  But this code is rather expensive.  Here are some
results (similar to what Jim Meyering posted before).

The test creates a deep directory hierarchy on a tmpfs filesystem.  Then
rm -fr is used to remove all directories.  Without syscall support I get
this:

real    0m31.921s
user    0m0.688s
sys     0m31.234s

With syscall support the results are much better:

real    0m20.699s
user    0m0.536s
sys     0m20.149s
=====



> We'll need Signed-off_by:'s for all these patches.

I'm not sure how you want to handle this.  I don't particularly want to
resent all of them just for this purpose.

[PATCH 1/3] updated *at function patch

Signed-off-by: Ulrich Drepper <drepper@redhat.com>

[PATCH 2/3] updated *at function patch

Signed-off-by: Ulrich Drepper <drepper@redhat.com>

[PATCH 3/3] updated *at function patch

Signed-off-by: Ulrich Drepper <drepper@redhat.com>


-- 
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 251 bytes --]